# Account settings

Update your name, display currency, app timezone, and password, and learn which changes require support.

Canonical: https://docs.equalai.io/account/account-settings
Markdown: https://docs.equalai.io/account/account-settings.md

The **Settings** page contains **Profile Settings** and **Security** for your signed-in account. These are personal settings, not the selected project's brand identity or the workspace's subscription configuration.

## Update Profile Settings

1. Open your account's **Settings** page.
2. Review **Full name**. It must not be blank.
3. Choose **Display currency** and **App timezone** as needed.
4. Choose **Save**.
5. Wait for the save to complete before navigating away.

### Email

The **Email** field shows the account email address. You cannot update it directly, even if the input lets you type a different address. Contact [support](https://docs.equalai.io/account/support) for an email-change request rather than repeatedly trying to save it here.

If you were invited to a project, an email mismatch can also mean you signed in with a different account. Check the intended invitation address with your workspace admin before requesting a change.

### Display currency

**Display currency** is a presentation preference. It does not change the currency of an existing subscription or rewrite a charge. For the authoritative transaction amount and currency, review the billing or checkout confirmation.

To manage a plan, renewal, or payment issue, a workspace owner/admin should use [Billing](https://docs.equalai.io/account/billing).

### App timezone

**App timezone** sets the timezone preference used by supported date and schedule displays. It is separate from a project's **Target Country** and **Target Language**.

Changing this preference is not the same as changing the stored tracking schedule. Open the [schedule dialog](https://docs.equalai.io/tracking/schedules), check its displayed timezone and upcoming-run preview, and save any intentional schedule change separately.

Do not assume every historical date label is reformatted identically. Check the timezone shown in the view when comparing timestamps or coordinating a run with another teammate.

## Change or set a password

In **Security**, the controls depend on whether the account already has a password.

### If a password is already configured

1. Enter **Current Password**.
2. Enter **New Password**.
3. Enter the same value in **Confirm New Password**.
4. Choose **Change password**.

The new password must contain at least six characters. Use a unique, stronger password rather than treating the minimum length as a recommendation. The confirmation must match before the change can be submitted.

### If the account has no password yet

The page explains that you can set one for future email-and-password sign-in. Enter **New Password** and **Confirm New Password**, then choose **Set your password**. There is no Current Password field in this state.

A successful change clears the password inputs. Use the new password on your next sign-in. Never send the old or new password to support.

## Troubleshooting

- **Profile Save disabled:** ensure Full name contains text and wait for any pending request to finish.
- **Password action disabled:** check the minimum length and matching confirmation; provide the current password when required.
- **Current password rejected:** verify the credential before retrying. Do not reuse a different provider's password.
- **Email change fails:** direct changes are not supported; contact support.
- **Wrong project name:** edit [Project settings](https://docs.equalai.io/account/project-settings), not Full name.
- **Timezone looks wrong on a schedule:** review the timezone and next-run preview in the scheduling dialog rather than assuming that changing a profile preference rescheduled work.

## Related guides

- [Project settings](https://docs.equalai.io/account/project-settings)
- [Billing](https://docs.equalai.io/account/billing)
- [AI Assistant Access](https://docs.equalai.io/account/ai-assistant-access)
- [Support](https://docs.equalai.io/account/support)


---

# AI Assistant Access

Connect an assistant to approved EqualAI projects, revoke read-only access, and understand the separate external-provider connection workflow.

Canonical: https://docs.equalai.io/account/ai-assistant-access
Markdown: https://docs.equalai.io/account/ai-assistant-access.md

**AI Assistant Access** lets an AI assistant read EqualAI information from projects you approve. It uses MCP, a connection format supported by compatible assistants. You sign in to EqualAI and approve the project scope in your browser.

This is different from connecting an external provider for EqualAI to call. The two directions have separate connections and confirmations.

## What an assistant can access

An approved assistant can read supported project information such as project details, tracked prompts, competitors, metrics, citation/report metadata, recommendation titles, and some recent answer samples. Read-only does not mean public: approve only an assistant with which you intend to share that project information.

The connection cannot create, edit, or delete EqualAI records; manage billing or team access; run prompts; or spend EqualAI credits.

Access remains limited by your current project permissions and usable subscription access. A prior approval does not preserve access after you lose permission to the project.

## Ask for product help

The same connection can search EqualAI's documentation and read complete guides. No second connector is needed. Try asking, “How does EqualAI calculate AI visibility?” or “Where do I change a prompt's tracking frequency?”

Documentation explains how the product works; project results show what happened in your workspace. An assistant can use both, but a guide is not evidence of your project's performance. Documentation citations open public pages on **docs.equalai.io**, unlike the private record links described below.

The assistant reads a bundled copy of the guides, updated with application releases, rather than fetching the website live. If a guide and the assistant disagree, open the cited page for the latest published instructions. The MCP connection still requires valid authorization and access to at least one approved project, even when asking only for documentation.

## Connect an assistant

1. Sign in and open **AI Assistant Access**.
2. Under **Connect an assistant**, use **Copy** beside the displayed server URL.
3. Choose the **Claude**, **ChatGPT**, **Claude Code**, or **Other assistants** instructions that match your client.
4. Add the displayed server URL in the assistant's connection settings and begin authentication.
5. In the EqualAI authorization page, inspect the client information and select only the projects you intend to share.
6. Approve the request, then return to the assistant to finish connecting.
7. Check **Connected assistants** for the completed connection and its project badges.

![AI Assistant Access setup showing the public MCP server URL and assistant instruction tabs with Claude selected](https://docs.equalai.io/images/guides/mcp.webp)

***1.** Copy the public MCP server URL. **2.** Choose your assistant for its setup steps. The Claude instructions are shown here; this setup view does not indicate that an assistant has connected or received project access.*

Use the URL displayed by EqualAI rather than a guessed endpoint. For Claude Code, copy the command shown in its tab rather than substituting an unrelated API address. Assistant menus and connector availability can vary by product and plan.

An authorization request expires after ten minutes. Up to 20 projects can be approved in one request. If the request expires, start the connection again from the assistant.

### Check the client before approving

Client names are self-reported, not proof of identity. Only approve a connection you initiated. Inspect the destination and project list; cancel an unexpected request rather than treating a familiar assistant name as verification.

No available project can mean you lack current project access or a qualifying active subscription. Ask an authorized workspace admin to review that access instead of approving a different project to work around it.

## Review and revoke connections

**Connected assistants** shows active connections, approved projects, and the connection date.

To end one:

1. Choose **Revoke access** for the intended connection.
2. Review the confirmation.
3. Choose **Confirm revoke**.

Revocation stops future access through that connection. It cannot retrieve information already shared with the assistant or delete the assistant's existing conversation history.

Approved attempts that the assistant never completed grant no access. They clear after 30 minutes; use **Remove** to clear one sooner, then reconnect from the assistant if needed. The ended-connection list distinguishes **Revoked** from **Not completed**. The page shows the newest 200 connections.

## Open a linked EqualAI record

An assistant may return a link to a read-only EqualAI record page for a project, tracked prompt, competitor, report, citation, or recommendation.

You must sign in and still have access to the associated project. A link does not make the record public or give its recipient access. Missing or inaccessible records may not be shown.

These browser pages display selected fields; report payloads and captured answer bodies are not included on the record page. Use **Open dashboard** to continue in EqualAI or **Manage assistant access** to review connections. HTTPS source links can take you to an external website.

## External MCP connections: the other direction

**External MCP connections** lets EqualAI send a manually confirmed, read-only tool request to a reviewed external provider. It does not give that provider general access to your EqualAI projects and is not the same as an inbound assistant grant.

### Connect a reviewed provider

A workspace owner/admin must connect or disconnect a provider.

1. Select the intended accessible project.
2. Choose **Connect a provider** and a **Reviewed provider**.
3. Review the project credential-sharing acknowledgement.
4. Complete the displayed OAuth flow or enter the requested bearer token.
5. For a bearer connection, choose **Save encrypted bearer connection**.
6. Verify that the connection shows **Connected**, rather than **Needs attention**.

The credential is shared for authorized use within that project. Authorized project users, including viewers, can confirm reviewed tool calls without seeing the stored credential. Do not connect a personal credential unless its use by those project users is acceptable.

If there are no reviewed providers or no accessible active projects, there may be nothing available to connect. Arbitrary custom endpoints are not an option in this workflow.

### Review and confirm a tool call

1. Select a reviewed tool from the connection.
2. Read its **Reviewed JSON input schema**.
3. Enter the **Exact JSON arguments** you intend to send.
4. Choose **Review disclosure**. This prepares a review; it does not send the tool request yet.
5. Inspect **Confirm external disclosure**, including the provider endpoint, tool, and arguments.
6. Choose **Confirm and send these exact arguments**, or **Cancel**.

The confirmation lasts two minutes and can be used once. If it expires, review the disclosure again. Changing the intended arguments requires a new review rather than reusing a previous confirmation.

Provider discovery sends the stored credential and lists metadata; it does not fetch provider resources. Tool responses are shown as **Untrusted provider output**. Treat them as third-party data, not instructions to run code, change settings, or disclose more information. Links, HTML, and instructions in that output are not automatically executed.

The workflow does not automatically chain tools, import provider resources, generate content, or make payments.

### Disconnect and recover

Disconnecting removes the EqualAI connection, but does not revoke the credential at the provider. Revoke or rotate it with the provider separately when appropriate.

- **Needs attention:** ask the workspace admin to review authorization and reconnect if needed.
- **Expired confirmation:** select the tool and review the disclosure again.
- **Changed tool schema:** the reviewed configuration must be updated before it can be used; contact support.
- **Rate limit:** wait one minute before retrying.

## Related guides

- [Team roles and project scope](https://docs.equalai.io/account/team-members)
- [Workspace subscription](https://docs.equalai.io/account/billing)
- [Support](https://docs.equalai.io/account/support)


---

# Billing and workspace usage

Review subscription status, shared capacity and credits, plan changes, and renewal controls without confusing workspace billing with project usage.

Canonical: https://docs.equalai.io/account/billing
Markdown: https://docs.equalai.io/account/billing.md

The **Subscription** page manages billing for the workspace. Opening it from a selected project's settings does not make the subscription project-specific.

Workspace owners and admins can access billing and perform subscription actions. A member, viewer, or project-only administrator should ask a workspace admin to review billing restrictions.

## Read subscription status

The page can show **Active**, **Past due**, **Cancelled**, or **No subscription**. Use the current status and billing-period information when deciding whether a payment or renewal action is needed.

Choose **Refresh** after returning from checkout or resolving a payment issue. A pending update can take time to appear; repeatedly opening checkout is not a reliable way to refresh it.

## Understand capacity and execution credits

| Item | Scope | What it represents |
| --- | --- | --- |
| Project slots / Project credits | Shared across workspace projects | Capacity to maintain projects. |
| Prompt slots / Prompt credits | Shared across workspace projects | Capacity for tracked prompts. |
| AI response credits | Shared across workspace projects | Budget consumed by prompt execution. |
| Competitor limit | Per project | Number of competitors tracked within each project. |

The words “prompt credits” can refer to tracked-prompt capacity on the plan page. They are not interchangeable with the AI credits spent on a run.

Usage for a selected project helps explain where capacity is being used; it does not give that project a separate credit balance. Work queued from another project can reduce the workspace balance available to your current project.

Read the reset information shown for the current credit allowance. Some allowances reset on the displayed interval; do not assume all purchased or included credits refresh on the same day or roll over.

## Compare plans and add-ons

Choose **View Plans** to review current prices, billing intervals, limits, and model/add-on availability. Available options can change with your current subscription, so use the page's values rather than an old screenshot or a previously quoted price.

Before checkout, confirm:

1. The intended workspace.
2. The plan and monthly or annual billing interval.
3. Included project/prompt capacity and AI credits.
4. Any model or other add-on selection.
5. The amount, currency, and timing shown in the final confirmation.

A model shown in the scheduling dialog may remain locked until its required plan or add-on is active. Choosing it elsewhere is not a substitute for activation.

**Display currency** in account settings is only a display preference; it does not change the subscription's billing currency.

## Switch billing interval

When **Switch to annual** is offered, review the change confirmation before proceeding. The change can take effect immediately and involve proration. Rely on the amount and effective timing shown by the billing flow; do not assume the change waits until the next monthly renewal or guarantees a refund.

## Cancel or resume renewal

For an active subscription, **Cancel renewal** schedules cancellation at the end of the current billing period. It is not an instruction to delete the workspace or an immediate refund request.

If cancellation is scheduled and **Resume renewal** is offered, use it to keep renewal enabled. Verify the updated status and period information afterward.

A cancelled or inactive subscription can show activation options instead of active-plan controls. Review the displayed plan rather than assuming that an old plan is still purchasable or that a new trial begins.

## Budget tracking runs

A run's cost depends on the prompts, selected models, and fan-out. Instant Run adds its displayed premium multiplier. The schedule dialog estimates the run cost and available balance; the available balance can change if another teammate queues work.

Saving a schedule does not spend credits. Executing it does. A schedule can therefore remain saved while a later run is blocked by insufficient balance or unavailable model access.

See [Schedules](https://docs.equalai.io/tracking/schedules) for selected-prompt runs and the difference between Standard and Instant delivery.

## Troubleshooting

- **Cannot open billing:** ask a workspace owner/admin; project editing permission is not enough.
- **Plan is active but a project cannot be added:** review shared project capacity across all workspace projects.
- **Competitor limit reached:** inspect the competitor count in that project, not a workspace total.
- **Run blocked despite spare prompt slots:** check AI response credits; capacity and execution budget are separate.
- **Paid but status is stale:** use Refresh and retain the payment reference for support.
- **Payment issue persists:** contact support with the visible status and a redacted billing reference. Never share full payment-card details.

## Related guides

- [Project settings](https://docs.equalai.io/account/project-settings)
- [Team permissions](https://docs.equalai.io/account/team-members)
- [Schedules and credit estimates](https://docs.equalai.io/tracking/schedules)
- [Support](https://docs.equalai.io/account/support)


---

# Project settings

Configure project identity, targeting, brand matching, email preferences, and connected-data settings while keeping workspace billing separate.

Canonical: https://docs.equalai.io/account/project-settings
Markdown: https://docs.equalai.io/account/project-settings.md

**Project Settings** controls the selected project's identity and reporting defaults. Confirm the project before editing, especially if you manage several brands. Project administration requires an owner or admin; regular members can edit tracking content but cannot manage these settings.

The visible tabs are **General Information**, **Billing**, **Email Notifications**, **GA4 Conversions**, **GSC Data**, and **Reputation Dimensions**. Choose **Save changes** for the settings you edit and wait for confirmation before leaving.

## General Information

### Project Identity

| Field | What to use it for |
| --- | --- |
| Name | The brand name to recognize and display. |
| Url | The project's website. This cannot be changed here. |
| Chart colors | Visual identification in charts, not a change to scoring. |
| Description | A concise description that distinguishes this brand and its offering. |
| Industry | The relevant business category. |
| Target Country | The project's market default. |
| Target Language | The language default; review it after changing country. |

Changing country can update the suggested language. If your market uses another language, choose it explicitly before saving. These defaults do not translate old answers or rerun historical prompts.

Although **Url** is shown with the identity fields, the existing project URL is not editable through this form. If the project was created for the wrong website, contact support before making a replacement project that could consume additional capacity.

### Exact subdomain matching

This option controls owned-citation matching. With exact matching enabled, a cited host must match the configured project host exactly. Without it, the configured host and its child subdomains can count as owned.

**Synthetic example:** for the fictional host `example.com`, normal matching can include `help.example.com`. If the configured host is `shop.example.com`, normal matching does not make the sibling `help.example.com` the same host. Exact matching restricts the match to the configured host itself.

### Alternative Names

Add distinctive names used for the same brand: alternate spellings, former names, or recognizable acronyms. The project name and domain are matched already.

- Maximum 12 additional names.
- Each name must be 2–60 characters and no more than six words.
- Names under four characters must be uppercase 2–3-character acronyms.
- Domains, overly generic fragments, and redundant variants may be rejected.

Matching is case-insensitive and respects word boundaries. Prefer specific names over broad category terms. Review validation feedback rather than trying to pad the list. Competitor aliases are managed separately in [Competitors](https://docs.equalai.io/tracking/competitors).

## Billing

This tab concerns the workspace subscription, not a subscription owned by the selected project. Project slots, tracked-prompt capacity, and AI credits are shared across workspace projects; competitor limits apply per project.

Billing access and subscription actions require a workspace owner or admin. See [Billing](https://docs.equalai.io/account/billing) for cancellation, renewal, upgrades, and usage.

## Email Notifications

Use **Weekly Visibility Summary** to control that project's weekly visibility-summary preference. Save the change. This is not the control for your password, app timezone, or an AI model's run schedule.

If you need to change the tracking cadence rather than email preferences, use [Schedules](https://docs.equalai.io/tracking/schedules).

## GA4 Conversions

Review the Google Analytics connection status, **Primary Property**, and **Conversion Events**. Event names can be entered one per line or separated by commas.

Saving this event-name list does not filter the current GA4 traffic report. Do not use it as evidence that an event has been configured or marked as a key event inside Google Analytics.

**Disconnect Google Analytics** stops fetching new reports through the connection. Confirm the project and account before disconnecting. See [Google Analytics](https://docs.equalai.io/integrations/google-analytics) for the reporting workflow.

## GSC Data

Review the Search Console account and selected site. If disconnected, use **Connect Google Search Console**. **Remove GSC Data** stops loading new Search Console query data for prompt research; it is not a command to delete your website or its Google Search Console property.

See [Google Search Console](https://docs.equalai.io/integrations/google-search-console) for connecting a property and using imported queries.

## Reputation Dimensions

Choose the dimensions to use as defaults for future reputation and sentiment reporting. Changing these selections does not recalculate an earlier report. Choose dimensions that are meaningful for this project rather than assuming every category applies to every industry.

## Troubleshooting

- **Cannot save:** check that you are an owner/admin for the project and that subscription access is usable.
- **Alternative name rejected:** shorten it, remove a generic fragment, or check for a duplicate; do not repeat the domain.
- **Billing unavailable:** project access alone is insufficient for workspace billing administration.
- **Unexpected connected property:** verify the Google account and selected property/site before using its reports.
- **Old results unchanged:** changing defaults does not itself queue new tracked answers.

## Related guides

- [Personal account settings](https://docs.equalai.io/account/account-settings)
- [Team roles and project access](https://docs.equalai.io/account/team-members)
- [Scheduling](https://docs.equalai.io/tracking/schedules)
- [Support](https://docs.equalai.io/account/support)


---

# Get support

Choose a support channel, collect useful troubleshooting details, and share a report without exposing account or customer information.

Canonical: https://docs.equalai.io/account/support
Markdown: https://docs.equalai.io/account/support.md

The **Support** page provides chat, email, Help Center, and subscription shortcuts. Choose the route that fits the question; opening a support channel does not change your tracking or billing settings.

## Contact support

- Choose **Email support** to open an email to [support@equalai.io](mailto:support@equalai.io).
- In **Contact support**, choose **Open chat** to open the support chat widget.
- In **Help Center**, choose **Open Help Center** to browse the in-app help entry point.
- In **Subscription help**, choose **View subscription** to review the workspace plan and usage. Billing access requires a workspace owner/admin.

If chat does not open, use email directly. Chat may depend on browser script or network availability; there is no need to disable your organization's security controls to report an issue.

No fixed response time or priority-service guarantee is implied by these controls.

## Send a useful issue report

Include enough context to reproduce the problem without revealing sensitive data:

1. The page name and action you attempted.
2. Whether the issue affects one project or several.
3. What you expected and what happened instead.
4. The exact visible error message, with sensitive text removed.
5. Approximate time and timezone.
6. Whether refreshing or retrying once changed the outcome.
7. A cropped, redacted screenshot if it helps.

**Synthetic example:** “On the example project's schedule dialog, Save schedule returned an error after changing the weekly days. The same page reopened with the old days.” This communicates the workflow without exposing real prompts, results, or account identifiers.

For an execution problem, say whether the run was rejected, queued, still running, or completed with an error. A queued confirmation and a completed result are different states.

## Keep sensitive information out

Do not send passwords, session cookies, authorization links, assistant tokens, provider credentials, full payment-card details, or unredacted customer data. If the issue involves an integration or assistant connection, share the provider name and visible status—not the stored secret or the full authorization request URL.

For a billing issue, use a redacted payment reference and the visible subscription status. For an email change, start with a request for help and follow the support verification process rather than sharing credentials.

## Quick checks before contacting support

| Symptom | First check |
| --- | --- |
| A teammate cannot see a project | Correct sign-in email, accepted invitation, and project role. |
| A save action is unavailable | Required fields, pending requests, role, and plan access. |
| A prompt run cannot start | Eligible prompts, selected models, credits, and cooldown. |
| Subscription status looks stale | Choose Refresh after returning from checkout. |
| Imported tags are missing | CSV header, two plain-text columns, duplicates, and batch size. |
| An assistant cannot read a project | Approved project scope, current project access, and subscription access. |

## Related guides

- [Onboarding](https://docs.equalai.io/getting-started/onboarding)
- [Team members](https://docs.equalai.io/account/team-members)
- [Billing](https://docs.equalai.io/account/billing)
- [Schedules](https://docs.equalai.io/tracking/schedules)
- [AI Assistant Access](https://docs.equalai.io/account/ai-assistant-access)


---

# Team members and permissions

Invite teammates, manage pending invitations, and distinguish workspace-wide administrators from project-scoped members and viewers.

Canonical: https://docs.equalai.io/account/team-members
Markdown: https://docs.equalai.io/account/team-members.md

**Team Members** shows the people who can access the selected project, including workspace owners and admins. Before inviting someone or changing access, confirm both the selected project and the intended role.

## Understand role scope

| Role | Access and responsibilities |
| --- | --- |
| Owner | Workspace-wide project access and administration. Owner entries do not have role or removal controls on this page. |
| Admin | Workspace admins can access every project and manage workspace billing and team access. |
| Member | Can view and edit the project they were invited to; cannot manage project settings, billing, or team access. |
| Viewer | Read-only access to the project they were invited to. |

Inviting an **Admin** grants workspace-wide access. Inviting a **Member** or **Viewer** grants access only to the selected project. Those invitations do not give access to unrelated projects.

Only workspace owners/admins can grant workspace admin access. A project-only administrator does not automatically have authority to manage the entire workspace's billing or administrators.

## Invite a member

1. Open **Team Members** for the intended project.
2. Choose **Invite member**.
3. Enter **Email address**.
4. Choose **Role**, checking the scope above.
5. Choose **Send invitation**.

The invitation appears in the pending list until accepted, revoked, or expired. It is not the same as an active member. Invitations expire after seven days.

**Synthetic example:** invite a fictional analyst to the example project as **Viewer** if they only need to review its results. Choose **Member** if they should maintain prompts and competitors. Do not choose Admin merely to avoid a missing project permission.

You cannot invite a new Owner through this role selector.

## Manage pending invitations

- **Resend** sends a replacement invitation with a new expiry. Ask the recipient to use the latest message.
- **Revoke** withdraws a pending invitation. It does not remove someone who already accepted and has active membership.

A project-only administrator cannot manage workspace Admin invitations. If an invitation is unavailable to manage, ask a workspace owner/admin rather than sending a second higher-privilege invitation.

## Change a member's role

Use the member's role control, then review the result in the list. Role changes are access changes, not just labels.

- Promoting a person to workspace **Admin** gives access across the workspace.
- Demoting a workspace member to project **Member** or **Viewer** replaces workspace-wide access with the selected project's access.
- You cannot change your own role here.
- Owner entries do not expose role changes on this page; do not use it as an ownership-transfer workflow.

Confirm the intended outcome with the team before using demotion to restrict access. For billing administration, the relevant permission is workspace owner/admin—not simply access to a project.

## Remove a member

Use the removal action and confirm the intended person and scope:

- Removing a project-only member removes that person's access to the selected project.
- Removing a workspace member removes their workspace membership and project memberships throughout that workspace.

You cannot remove yourself using this control. Owner entries have no removal action on this page.

If the person also connected an AI assistant, future data access is checked against their current permissions. Review [AI Assistant Access](https://docs.equalai.io/account/ai-assistant-access) when ending an assistant connection; revoking access cannot retrieve information already shared.

## Troubleshooting

- **Missing project after acceptance:** confirm the account email and that the invitation targeted the correct project.
- **Pending invite not working:** check whether seven days have elapsed, then resend and use the latest link.
- **Admin option unavailable:** the current account may not have workspace administration rights.
- **Cannot change your own role:** ask another authorized owner/admin.
- **Viewer cannot save:** read-only access is intentional; an authorized admin must change the role if editing is needed.
- **Unexpected users in a project:** workspace owners/admins have access to all workspace projects and can appear alongside project members.

## Related guides

- [Project settings](https://docs.equalai.io/account/project-settings)
- [Workspace billing](https://docs.equalai.io/account/billing)
- [Account settings](https://docs.equalai.io/account/account-settings)
- [Support](https://docs.equalai.io/account/support)


---

# AI Agents

Inspect recognized crawler user agents in Cloudflare analytics or preview a local server log.

Canonical: https://docs.equalai.io/ai-visibility/ai-agents
Markdown: https://docs.equalai.io/ai-visibility/ai-agents.md

AI Agents shows requests identified by user-agent text. Use it to investigate crawler activity, not to measure human visitors, citations, or whether an AI company trained on your pages. For attributed website sessions, use [Traffic Analytics](https://docs.equalai.io/ai-visibility/traffic-analytics).

## Connect the project

Open **AI Visibility → AI Agents**. The page heading is **Agent Analytics**. The connection workflow has three stages: **1. Connect Cloudflare**, **2. Select zone**, and **3. Analytics**. A workspace administrator needs a token with access to the intended Cloudflare zone. Follow [Connect Cloudflare](https://docs.equalai.io/integrations/cloudflare) for permissions and setup.

![Agent Analytics before connection, showing the Cloudflare setup stages and optional local server-log preview](https://docs.equalai.io/images/guides/agents.webp)

*EarlySEO's unconfigured connection screen, with an empty token field. **1.** Connection, zone selection and analytics. **2.** Create a limited Cloudflare API token. **3.** Optional local log preview. No connected-zone request data is shown; this is not evidence of zero crawler activity.*

## Read the analytics

1. Confirm the **Connection** and **Zone** before interpreting results.
2. Select **7D**, **30D**, or **90D**. The default is 7D.
3. Check **Data** and **Updated** for availability and freshness.
4. Compare **Total vs AI requests**, then use the bot chips in **Bot request trends** to show or hide individual lines. Up to six bot series are offered in that chart.
5. Review **Bots in view** for counts by recognized bot family.

| Card | Interpretation |
| --- | --- |
| Total requests | All requests returned for the zone and period, not only bot requests. |
| AI bot requests | Requests in returned user-agent groups that match the recognized bot catalog. |
| Bots seen | Number of recognized bot families with matching activity, not the number of unique full user-agent strings. |
| Companies seen | Companies mapped from those recognized families. |

The live catalog includes GPTBot, ChatGPT-User, OAI-SearchBot, ClaudeBot, Claude-SearchBot, PerplexityBot, Perplexity-User, Googlebot, GoogleOther, Applebot, Bingbot, and Meta-ExternalAgent. Some are general search crawlers. The **AI bot requests** label is therefore broader than exclusively AI-assistant traffic.

## Understand scope and freshness

Cloudflare analytics is requested in daily, trailing 24-hour windows with UTC date labels. User-agent analysis uses up to the top 200 returned groups per window, so smaller groups can be absent. Matching a user-agent name is not authenticated proof of the caller's identity.

Analytics can be cached for 30 minutes. **Sync** requests updated connection and analytics data; a failed sync can leave older results visible. Check the status, update time, and any error before relying on the chart.

## Preview a local log

On the connection screen, **Analyze server logs** provides a separate local preview:

1. Select **Choose log file** and choose a `.csv`, `.log`, or `.txt` file no larger than 8 MiB.
2. Read the line count, **AI crawler hits**, **Bots detected**, and top matched bots.
3. Use the preview to decide whether a fuller investigation of your server logs is needed.

The file is processed in your browser and is not uploaded or saved as a Cloudflare integration. The preview searches nonempty lines for bot-name patterns; it does not parse timestamps, validate a log format, or apply the dashboard's date range. One line can match multiple bot families. Its recognition list differs from the live Cloudflare catalog, so the two totals need not agree.

## Resolve missing data

- No zones: confirm token permissions and zone scope, then use **Sync now** on the zone-selection screen.
- Request error: inspect the displayed Cloudflare error and reconnect or sync with an authorized token.
- No requests: verify the selected zone and range. An empty successful result differs from a failed request.
- Total requests but no AI bot trends: no recognized returned user-agent groups matched; this is not proof that no crawlers visited.

**Disconnect** removes EqualAI's stored connection credential, not the token at Cloudflare. See [connection lifecycle](https://docs.equalai.io/integrations/cloudflare#refresh-and-disconnect).


---

# AI Visibility Reports

Interpret the current tracked-response visibility summary, its citation and sentiment context, and separate saved-report actions.

Canonical: https://docs.equalai.io/ai-visibility/ai-visibility-reports
Markdown: https://docs.equalai.io/ai-visibility/ai-visibility-reports.md

**AI Visibility Reports** summarizes analytics already available for the current project. Open it from the Optimization navigation or choose **AI Visibility Report** in **GEO Optimization**. Opening this page does not run prompts, query models again, or analyze your website's HTML.

## Read the visibility score

The score is the rounded percentage of responses in the loaded analytics set that contain a tracked brand mention. The denominator is responses, not unique prompts or competing brands. It is not share of voice and does not measure what fraction of all AI answers on the internet mention you.

| Score | Rating |
| --- | --- |
| 70–100 | Strong |
| 50–69 | Average |
| 30–49 | Below Average |
| 0–29 | Weak |

When no responses are available, the calculated score is zero. Treat missing analytics as a data-availability issue rather than a measured market position. The report has no independent date or model selector; its summary uses the loaded project analytics, not filters you may have applied elsewhere.

## Review the sections

- **Known For** lists up to five distinct prompt topics, excluding General and Others. These are tracked-topic labels, not a model-derived explanation of your reputation.
- **Quick Wins**, **Key Strengths**, and **Improvement Opportunities** give reviewable guidance. Confirm that an item applies before changing your content.
- **Brand Reputation** summarizes sentiment among brand-mentioned responses. Unknown or missing sentiment is treated as Neutral for this summary; the most frequent category is shown.
- **Citation Examples** shows up to three leading citation hosts across the loaded citations. A host's displayed sentiment follows the overall brand sentiment, not a separate sentiment analysis of that website.
- **Sample Search Prompts** shows a limited selection of tracked prompts, not the complete prompt inventory. The summary uses up to seven prompt examples and four competitor entries.

The citation total uses owned-domain citations when any are available; otherwise it falls back to all loaded citations. Do not assume every displayed citation count or example refers to your website. Inspect individual response and citation evidence before making that claim.

## Decide what to do next

1. Confirm that the project has tracked prompts and completed response data.
2. Read the score with its response coverage in mind. Changing the set of tracked prompts can change the result even if your website is unchanged.
3. Investigate representative prompts and cited sources rather than relying only on the headline score.
4. Use [Prompt Explorer](https://docs.equalai.io/ai-visibility/prompt-explorer) to improve prompt coverage or [Optimization](https://docs.equalai.io/ai-visibility/optimization) to investigate technical website signals.

The assembled report time is not proof that a new model run occurred. To obtain new responses, use your prompt-tracking workflow rather than reopening this page.

## Saved Report History, Share, and Export

The page can also show **Saved Report History**. Its **Share** and **Export** actions depend on an available saved report and target that saved report, not a newly captured copy of the current live summary. They are disabled when no saved report is available. A saved artifact can therefore differ from the summary currently displayed.

If you see **No analytics yet**, first establish tracked-response data for the selected project. Do not run website checks expecting them to populate this response-based summary.


---

# Content Readiness

Review text length, heading structure, lists, tables, and authorship signals in the initial page HTML.

Canonical: https://docs.equalai.io/ai-visibility/content-readiness
Markdown: https://docs.equalai.io/ai-visibility/content-readiness.md

Content Readiness checks how the retrieved page is structured. It does not grade factual accuracy, originality, usefulness, or the likelihood that a model will quote it. The Optimization card is labeled **Content** and its action is **Check Content**.

## Review the page

1. Open **GEO Optimization → Check Content** for the correct project URL.
2. Inspect word count, H1/H2/H3 counts, lists, and question headings.
3. Review paragraphs, tables, **Has date**, **Has author**, and the heading outline.
4. Compare these signals with what the page needs to communicate, then make targeted editorial or template changes.
5. Publish changes and rerun from Optimization to save a new result.

The heading outline shows up to the first 12 headings. Text is extracted after scripts, styles, and tags are removed; it is not limited to a semantically isolated article body. Navigation, footer, and other page text can influence counts. Client-rendered content may be absent.

## Scoring rules

| Signal | Points |
| --- | --- |
| At least 300 words | 20 |
| Exactly one H1 | 10 |
| At least two H2 headings | 15 |
| At least one list | 10 |
| At least one question-style H1–H4 heading ending in a question mark | 20 |
| At least one table | 10 |
| A detected date or author signal | 15 |

Date detection uses a published-time metadata signal or a time element. Author detection uses supported author metadata or an author relationship. A visible byline or date in arbitrary text may not be recognized, and detecting a date does not establish that the content is current.

## Act on gaps without padding content

Use headings to describe the actual page, lists for genuinely scannable steps, and tables where a comparison is useful. Do not add irrelevant FAQs or unnecessary words just to pass a check. Some short landing pages or transactional pages can be effective without matching every threshold.

If the report disagrees with the browser, inspect the initial HTML and the project's target URL. For structural problems, also review [Site Structure](https://docs.equalai.io/ai-visibility/site-structure); for missing content in the response, use [Crawlability](https://docs.equalai.io/ai-visibility/crawlability).

## Report controls

**Refresh** obtains a new displayed analysis of the current project URL. **Export** downloads that analysis as JSON, and **Share** copies the report page link without granting access. Refreshing an open saved report does not overwrite its stored history payload; run **Check Content** from Optimization to retain another comparison point.


---

# Crawlability

Check what a non-JavaScript fetch can read from the project page and investigate missing crawler signals.

Canonical: https://docs.equalai.io/ai-visibility/crawlability
Markdown: https://docs.equalai.io/ai-visibility/crawlability.md

Crawlability examines the initial HTML returned for your project URL. It helps identify inaccessible content and missing page metadata before you investigate AI visibility. It does not impersonate every crawler, render JavaScript, or establish whether a search engine has indexed the page.

## Run and inspect the report

1. Verify the project's website domain.
2. Open **GEO Optimization** and choose **Check Crawlability**.
3. Review **Crawler Signal Scorecard**, **HTML Weight**, **Metadata Coverage**, and **Content Structure**.
4. Use **Recommendations** to identify changes worth making on the analyzed page.
5. Rerun from Optimization after publishing changes to save a new result.

A successful fetch is a prerequisite for meaningful page analysis. A fetch failure can return a zero/error result; do not treat that as proof that every individual page signal is missing.

## What the score measures

The score adds the weights of passed checks:

| Check | Points |
| --- | --- |
| Successful HTTP response | 25 |
| Title present | 10 |
| Meta description present | 10 |
| Canonical link present | 10 |
| Primary H1 present | 10 |
| JSON-LD block present | 10 |
| HTML size between 2,000 and 2,000,000 bytes | 10 |
| At least one crawlable link | 10 |
| Image alternative-text coverage of at least 80% | 5 |

These are presence and structural checks. For example, a JSON-LD block can satisfy presence without being semantically correct; use [Schema](https://docs.equalai.io/ai-visibility/schema) for parsing and type coverage. An HTML-size threshold is a diagnostic rule, not a page-speed measurement.

## Act on a result

- If the fetch fails, check the page's response, authentication requirements, and blocking rules before rewriting content.
- If text, headings, or links are missing, compare the server-returned HTML with the rendered browser page. Content supplied only by client-side JavaScript is outside this check's view.
- If metadata is absent, add an accurate title, description, and canonical in your website system, then verify the published HTML.
- Review image alternatives for usefulness and accessibility, not just a higher coverage percentage.

A passing report does not guarantee that a bot is permitted by robots.txt. Review [Robots.txt](https://docs.equalai.io/ai-visibility/robots-txt) separately.

## Keep and compare evidence

The report's analyzed time identifies its fetch. **Export** downloads the displayed JSON analysis. **Share** copies an access-controlled page link. **Refresh** displays a new analysis but does not rewrite a saved history entry; use Optimization for a new stored run. See [report lifecycle](https://docs.equalai.io/ai-visibility/optimization#saved-reports-refresh-export-and-share).


---

# Discoverability

Inspect basic indexing hints, page metadata, and root discovery files without mistaking them for confirmed indexing.

Canonical: https://docs.equalai.io/ai-visibility/discoverability
Markdown: https://docs.equalai.io/ai-visibility/discoverability.md

Discoverability combines signals from your project's initial page HTML with its root robots.txt, sitemap.xml, and llms.txt. It is a diagnostic checklist, not a search-engine indexing report or a full crawl-permission test.

## Run the check

1. Open **GEO Optimization → Check Discoverability**.
2. Review the indexing and file signals, then inspect individual passed and failed checks.
3. Confirm which missing signals are relevant to your page and publishing policy.
4. Publish justified fixes in your website, then rerun from Optimization to save a new result.

## What contributes to the score?

| Signal | Points |
| --- | --- |
| No simple global robots.txt block detected | 20 |
| No meta robots noindex detected | 15 |
| Both title and meta description present | 15 |
| Canonical link present | 10 |
| Sitemap signal found | 15 |
| At least three Open Graph tags | 10 |
| JSON-LD block present | 10 |
| llms.txt found | 5 |

An **Indexable** result means the page did not contain the checked meta noindex signal. It does not verify that a search engine indexed the page, account for every response-header directive, or establish every bot's access.

The robots check looks for a simple global-block pattern. It is not the dedicated robots classifier. Use [Robots.txt](https://docs.equalai.io/ai-visibility/robots-txt) to inspect matched agents and restrictions. Similarly, JSON-LD presence here does not validate its syntax or properties; use [Schema](https://docs.equalai.io/ai-visibility/schema).

## Investigate a low score

- If discovery files are missing, verify their exact root locations and whether the server returns their actual content.
- If metadata is absent, inspect the initial HTML rather than only the browser's final DOM.
- If noindex or a robots restriction is intentional, retain it. An internal, staging, or deliberately excluded page should not be opened to crawlers to improve this score.
- If a sitemap exists elsewhere, this root-path check may not find it. Review the site's published configuration rather than assuming no sitemap exists anywhere.

The tool does not recursively crawl the sitemap, test every linked URL, or submit pages to search engines. Open Graph counts do not validate the content of every tag.

## Recheck and keep evidence

The analyzed time identifies the report's fetch. **Refresh** requests a new displayed analysis, **Export** downloads JSON, and **Share** copies the current access-controlled report URL. Run the check from Optimization to create another saved history result.

Continue to [Site Structure](https://docs.equalai.io/ai-visibility/site-structure) for link and heading signals or [llms.txt](https://docs.equalai.io/ai-visibility/llms-txt) for file-format checks.


---

# llms.txt

Inspect published llms.txt files, heading and link signals, and the optional full-content file.

Canonical: https://docs.equalai.io/ai-visibility/llms-txt
Markdown: https://docs.equalai.io/ai-visibility/llms-txt.md

This check inspects your website's root `/llms.txt` and `/llms-full.txt` files. Use it to verify what is already published, not to create a file inside EqualAI or prove that an assistant consumes it.

## Check published files

1. Open **GEO Optimization** for the intended project.
2. Choose **Generate llms.txt**. Despite this action label, the current tool analyzes the website's files; it does not generate, download a newly authored file, or deploy one.
3. Review **Status**, **Size**, **Sections**, **Links**, and **Checks**.
4. Read **llms.txt Content**, when available, to confirm that the fetched version is the one you intended to publish.

The files are fetched from the site's root, even if the project's page URL contains a deeper path.

## Interpret the score

| Signal | Points |
| --- | --- |
| Successfully fetched, nonempty llms.txt | 45 |
| At least one Markdown heading at levels 1–3 | 15 |
| At least one Markdown link | 25 |
| Successfully fetched, nonempty llms-full.txt | 15 |

If the main llms.txt is absent, the analyzer's score is zero even if a full-content file exists. History can present a missing file without a numeric score rather than as a completed quality assessment.

Sections and links are counted using text patterns. The check does not open every linked destination, assess the accuracy of your descriptions, or certify conformance with every interpretation of the llms.txt convention. A high score is not proof of AI indexing or citation.

## Publish or improve your file

Create or edit the file in your website's publishing system. Describe the site clearly, organize useful public resources with Markdown headings, and link to their canonical locations. Include only information intended to be public. An optional llms-full.txt can expose longer material if that fits your publishing policy.

After deployment, verify that the root file is publicly reachable and rerun the check from Optimization to save a new result. EqualAI does not synchronize edits made in a local file or upload content to your host.

## Resolve an unexpected result

- Missing file: check the exact root path, server response, and whether access restrictions or anti-bot challenges prevented retrieval.
- Zero sections or links: inspect the fetched text for recognizable Markdown headings and links rather than only an HTML error page.
- Old content: compare the analyzed time and use **Refresh** for a new displayed analysis. Rerun from Optimization when you need a new saved history entry.

**Export** downloads analysis JSON, not a newly generated llms.txt. **Share** copies an access-controlled report link. See [Optimization report controls](https://docs.equalai.io/ai-visibility/optimization#saved-reports-refresh-export-and-share).


---

# Optimization

Run technical checks for your project URL, interpret saved reports, and distinguish website diagnostics from tracked-response visibility.

Canonical: https://docs.equalai.io/ai-visibility/optimization
Markdown: https://docs.equalai.io/ai-visibility/optimization.md

Open **AI Visibility → Optimization** to reach **GEO Optimization**. This workspace combines eight technical checks with an AI Visibility summary. Technical scores describe observable website signals; they do not predict rankings or guarantee inclusion in an AI answer.

## Before running checks

Confirm the current project and its website domain in Project Settings. A missing domain prevents website analysis. Running checks requires project editing permission.

The analyzers retrieve the project URL's initial HTML and relevant root files. They do not execute your JavaScript application, crawl every linked page, or audit an entire product catalog. A page whose important content appears only after JavaScript runs may therefore look incomplete.

## Run a check

1. Open a check card or **Run a single check**.
2. Choose the corresponding action below.
3. Wait for the result, then inspect the evidence and recommendations in the report.
4. Make justified changes in your own website or publishing system. EqualAI does not apply those website changes for you.
5. Run the check again from Optimization to create a new history result for comparison.

| Action | What it examines |
| --- | --- |
| Generate llms.txt | [Published llms.txt and llms-full.txt files](https://docs.equalai.io/ai-visibility/llms-txt). Despite the action's name, it currently analyzes files rather than generating or publishing one. |
| Analyze robots.txt | [Crawler policy categories](https://docs.equalai.io/ai-visibility/robots-txt) from the root robots.txt file. |
| Check Crawlability | [Initial HTML accessibility signals](https://docs.equalai.io/ai-visibility/crawlability). |
| Analyze Schema | [JSON-LD presence, parsing, and selected types](https://docs.equalai.io/ai-visibility/schema). |
| Check Content | [Content Readiness](https://docs.equalai.io/ai-visibility/content-readiness): text length and page-structure signals. |
| Check Discoverability | [Metadata, indexing hints, and root discovery files](https://docs.equalai.io/ai-visibility/discoverability). |
| Check Site Structure | [Links, headings, canonical, and sitemap](https://docs.equalai.io/ai-visibility/site-structure). |
| Check Shopping Feed | [Product markup in the page](https://docs.equalai.io/ai-visibility/shopping-feed), not a submitted merchant feed. |
| AI Visibility Report | [A summary of tracked responses](https://docs.equalai.io/ai-visibility/ai-visibility-reports), not a new technical crawl or model run. |

**Run all checks** runs the eight technical checks. It does not run AI Visibility or execute tracked prompts.

![GEO Optimization entry screen with all-check and single-check actions and not-run check statuses marked](https://docs.equalai.io/images/guides/optimization.webp)

*EarlySEO before technical checks have run. **1.** Run all checks or choose one. **2.** Read the current status of each check. Not run yet is not a failed check or a score of zero; choosing a run action starts analysis.*

## Use Run history

The summary cards show **Reports generated**, **Completed**, **Needs attention**, and **Average score**. Check cards show the latest available result or **Not run yet**.

Use the report-type tabs, **All statuses**, and **Search by domain, check, or result** to find an earlier result. Available statuses include Completed, Failed, Insufficient data, Pending, and Processing. A failed or insufficient result is not equivalent to a measured zero score. Missing robots.txt or llms.txt files can appear without a numerical history score.

Open a row to inspect its report. Where offered, export a stored report as JSON or delete that history entry. Deleting a report does not change your website. The AI Visibility history entry can be a current, derived summary rather than a deletable stored technical run.

![EarlySEO Optimization run history with check-type filters, search and status filters, and the completed AI Visibility entry marked](https://docs.equalai.io/images/guides/optimization-history.webp)

*EarlySEO example history. **1.** Filter by check type. **2.** Search and status filters. **3.** Read an existing completed report. Only AI Visibility is Completed here; this derived summary is not evidence that a technical website check has run.*

## Saved reports, Refresh, Export, and Share

Opening a saved technical report displays its recorded payload. **Refresh** analyzes the current project URL and updates the displayed result; it does not replace that saved history payload. To retain a new history entry, run the check from Optimization.

On technical report pages, **Export** downloads the currently displayed analysis as JSON. **Share** copies the page URL; it does not make the report public or grant access to another person. **Back to GEO Optimization** returns to the relevant check type.

An automatically loaded technical result can be reused for 15 minutes. Explicit check runs and **Refresh** request new analysis.

## When a check cannot finish

Check the URL, public availability, and any displayed fetch error. Authentication requirements, anti-bot challenges, and server errors can prevent analysis. Some independently fetched files may still produce signals even if the main page cannot be read. Fix access deliberately; do not remove protections or intentional crawler restrictions merely to improve a score.

Scores use weighted checks rounded to a 0–100 result. Read the failed checks before prioritizing work: a lower score may reflect an intentionally absent or irrelevant feature rather than a defect.


---

# Prompt Explorer

Discover, draft, review, and create tracked prompts from your own questions, search data, discussions, or guided templates.

Canonical: https://docs.equalai.io/ai-visibility/prompt-explorer
Markdown: https://docs.equalai.io/ai-visibility/prompt-explorer.md

Choose **AI Visibility → Prompt Explorer**. The page heading is **Prompt Research**. This workflow creates prompts in the current project; generating drafts is not the same as executing those prompts against models.

## Start a workflow

1. Confirm the current project and choose **Get started** on a method card. A disconnected Google Search Console card can show **Connect & start** instead.
2. In **Target market**, select a country and language. These are single selections despite the plural labels **Target countries** and **Target languages**. Changing country resets the suggested language, so check both before continuing.
3. Review the project-derived brand, domain, topic, and industry context where offered, then choose **Continue**.
4. In **Configure source**, supply the method-specific inputs below and choose **Generate drafts**.
5. In **Review & create**, review and edit the proposed wording, select the prompts to keep, and remove unwanted drafts.
6. Check topic assignments, then choose **Create N prompts**. EqualAI returns to the tracked prompts page.

![Initial Add Manually Target market step with country and language selectors, prefilled EarlySEO brand context, and Continue marked](https://docs.equalai.io/images/guides/prompt-market.webp)

*EarlySEO example at the initial Target market step. **1.** Target countries and languages. **2.** Prefilled brand context. **3.** Continue to configure the source. Each market field accepts one selection; review the example defaults for your own project before continuing. No generated prompts are shown.*

Creating prompts requires project editing permission, an active usable subscription, and available workspace-wide prompt capacity. Draft estimates are not guarantees of the number saved: blank entries and duplicates can be removed, generation can return fewer suggestions, and creation can fail if an applicable limit is reached.

## Choose a source

| Method | Inputs and behavior |
| --- | --- |
| Add Manually | Enter a **Prompt**. The screen also offers an intent choice; newly created prompts from this workflow are currently saved with Research intent rather than preserving that choice. |
| Add in Bulk | Paste prompts one per line. **Questions only** keeps lines ending in `?`; **One per line** accepts the cleaned lines. Leading list markers and empty lines are removed. |
| Google Search Console | Connect an authorized Google account, select available query rows, and choose **Prompt style**: Question, Comparison, or Buyer. Queries become draft questions; this does not import historical AI responses. |
| People Also Ask (PAA) | Supply a **Seed query** and **Depth**: Standard or Deep. The workflow requests live search-question data and can use related keyword data and AI-assisted rewriting. Standard targets up to 8 questions; Deep targets up to 16. Results are not guaranteed to be verbatim PAA entries. |
| Reddit Insights | Search discussions, optionally narrow **Subreddit scope**, and select source discussions. Choose Pain points, Recommendations, or Objections. Drafts are constructed from the selected discussions; they are not a comprehensive analysis of Reddit sentiment. |
| Magic Prompts | Use brand/category context, **Audience**, and **Funnel stage**: Awareness, Evaluation, Purchase, or Retention. Current suggestions are generated from buyer-question templates, not a fresh model analysis of your website. |
| Topic Prompts | Enter **Topic or keyword** and select angles such as pricing, comparison, implementation, or alternatives. The method builds template-based variations. |
| Reputational Prompts | Choose dimensions such as Trust, Pricing, Customer support, Security, Accuracy, or Alternatives. The method creates questions to track; it does not itself measure reputation. |
| Translate Prompts | Paste one prompt per line, select the target country/language, and choose **Localized** or **Literal translation**. This method uses AI generation, so review the meaning, brands, and language before saving. |

The **Generate prompt ideas** group combines guided templates with AI translation. Treat all output as drafts to review, not verified evidence about your market.

![Prompt Research method cards for Google Search Console, People Also Ask, Reddit, and AI-assisted prompt sources](https://docs.equalai.io/images/guides/prompt-explorer.webp)

***1.** Search-data sources: GSC, PAA and Reddit. **2.** AI-assisted prompt sources. These cards are entry points, not generated results; the method table above distinguishes live source data, templates, and translation.*

## Work with source data

For Search Console, see [Google Search Console setup](https://docs.equalai.io/integrations/google-search-console). The query table includes clicks, impressions, and position. Its selected-query average position is an arithmetic average, not an impression-weighted aggregate. The connection's site determines the data; picking a target market does not add a country filter to that Search Console query request.

Reddit search offers sort choices Relevance, Hot, Top, New, and Most comments, plus time ranges from Past day through All time. Choose discussions that actually support the questions you want to track. Searching and selecting do not post anything to Reddit.

PAA, Reddit search, and translation depend on their relevant service availability. If the screen reports an unavailable service, retry later or use manual entry; do not treat an error as an empty research result.

## Review topics and duplicates

Drafts can be grouped into topics automatically. Read the topic shown with each prompt; grouping failure is not fatal and prompts can fall back to the workflow's topic. Small batches may remain under that shared topic. The Create action waits while grouping is in progress.

Edit wording before saving and use **Remove prompt** for unwanted drafts. Duplicate normalization during creation ignores case, repeated whitespace, and trailing question/exclamation/full-stop punctuation within the submitted batch. An exact-text match already in the current project reuses that prompt rather than creating another. Broader normalization does not apply to every existing prompt, and reusing one does not update its saved topic or locale.

After creation, use **Tracked prompts** to review the resulting inventory and continue the tracking workflow. Selecting a target locale records the prompt's locale; it does not prove that every template-based draft has been translated into that language.

## Resolve a blocked step

- **Generate drafts** unavailable: complete the required source input or select at least one query/discussion/dimension as appropriate.
- No Search Console queries: verify the connected site's access and available data; reconnect as a workspace administrator if required.
- Empty or unsuitable PAA suggestions: try a clearer seed, verify country/language, or enter reviewed prompts manually.
- Translation differs from the source meaning: edit or remove the draft before creating it.
- Creation error: inspect the message and check the tracked prompts list before retrying a large batch, so you do not unintentionally add duplicates.


---

# Robots.txt

Review how the published robots.txt is categorized for recognized crawlers without confusing policy with observed access.

Canonical: https://docs.equalai.io/ai-visibility/robots-txt
Markdown: https://docs.equalai.io/ai-visibility/robots-txt.md

The Robots.txt report retrieves the root robots.txt file for your project website and categorizes recognized user agents as **Allowed**, **Partial**, or **Blocked**. It evaluates a declared crawling policy, not actual request logs or an enforced security boundary.

## Analyze the file

1. From **GEO Optimization**, choose **Analyze robots.txt**.
2. Confirm whether the file was found and review the allowed, partial, and blocked counts.
3. Filter **AI Bots** by status to focus on relevant entries. Each card identifies its matched agent and any displayed blocked paths.
4. Review **Reference Bots** separately, then inspect **Recommendations**.
5. If present, inspect **robots.txt Content**. **Copy** copies that fetched text for review; it does not publish a replacement.

## Interpret the categories

The analyzer chooses the most specific matching user-agent token, merging groups of equal specificity. No matching group or no nonempty disallow rule results in Allowed. An unrestricted root disallow can result in Blocked; other nonempty disallow cases are categorized as Partial.

This is a summary classifier rather than an exhaustive URL-by-URL robots rules simulator. An **Allowed** card does not guarantee access to every URL, and a **Partial** card does not list every affected page. Inspect the source rules and test important URLs independently before changing policy.

If no file is found, the report shows default-allow categories and missing-file guidance. A network error or inability to retrieve a file is not proof that the site's access policy permits crawling. Also consider HTTP responses, authentication, firewalls, and meta or header directives.

## Score and scope

Only the AI bot catalog contributes to the access score. Each Allowed entry contributes 100, Partial contributes 50, and Blocked contributes 0; the average is rounded. **Reference Bots** do not contribute to that score.

A restrictive policy can be intentional. Do not allow unwanted crawlers just to increase the score. robots.txt is a public instruction for cooperating crawlers, not a way to protect private content.

## Make and verify a policy change

1. Decide which crawlers and paths should be permitted for your organization.
2. Edit the actual root robots.txt through your website's deployment or hosting process.
3. Confirm that the published file contains the intended rules and sitemap references.
4. Run **Analyze robots.txt** again to store a new result, then compare the bot categories and source text.

**Refresh** rechecks the current project website in the open report. **Export** downloads its displayed JSON. **Share** copies the current report URL without granting access. See [saved-report behavior](https://docs.equalai.io/ai-visibility/optimization#saved-reports-refresh-export-and-share).

Use [AI Agents](https://docs.equalai.io/ai-visibility/ai-agents) for observed user-agent activity and [Discoverability](https://docs.equalai.io/ai-visibility/discoverability) for additional indexing hints.


---

# Schema

Inspect JSON-LD parsing and selected structured-data types in the project's initial page HTML.

Canonical: https://docs.equalai.io/ai-visibility/schema
Markdown: https://docs.equalai.io/ai-visibility/schema.md

Schema checks structured data in the initial HTML returned for your project URL. It helps you find missing or malformed JSON-LD, but is not a complete schema.org validator or a guarantee of rich-result eligibility.

## Analyze structured data

1. Open **GEO Optimization → Analyze Schema**.
2. Review **JSON-LD blocks**, **Schema types**, **Invalid blocks**, and **Microdata items**.
3. Read the checks and recommended type coverage alongside the page's actual purpose.
4. Correct relevant markup in your website, publish it, then run the check again from Optimization.

A type is discovered from top-level objects, arrays, or `@graph` entries. Types buried in other arbitrary nested objects may not appear. Markup injected only after JavaScript runs is outside this initial-HTML analysis.

## Understand the score

| Signal | Points |
| --- | --- |
| At least one JSON-LD block | 30 |
| Organization or WebSite type | 25 |
| BreadcrumbList type | 15 |
| Article, Product, or FAQPage type | 20 |
| At least one JSON-LD block and all blocks parse as JSON | 10 |

Microdata items are counted separately; they do not substitute for these JSON-LD scoring checks. A block that parses as JSON can still have missing properties, inaccurate values, unsupported vocabulary, or unsuitable nesting. “Valid” in this context means parseable JSON, not complete semantic validation.

## Choose the right correction

- **Invalid blocks:** inspect JSON syntax, quoting, and trailing commas before adding more types.
- **No JSON-LD:** confirm whether your page intentionally uses another format or whether the markup exists only in the browser-rendered DOM.
- **Missing type coverage:** add only types that truthfully describe the page. Do not label unrelated content as a product, article, or FAQ solely to earn points.
- **Unexpected types:** inspect top-level arrays and graph entries and verify the correct URL was analyzed.

Check important structured data with appropriate external validators as well as the rendered and server-returned page. The report does not test every property or linked entity.

## Save and share findings

**Export** downloads the displayed analysis as JSON. **Share** copies the report page URL without making it public. **Refresh** reanalyzes the current project URL for the displayed view; running the check from Optimization creates a new history entry.

For commerce-specific signals, continue to [Shopping Feed](https://docs.equalai.io/ai-visibility/shopping-feed). For missing server-rendered content, review [Crawlability](https://docs.equalai.io/ai-visibility/crawlability).


---

# Shopping Feed

Check product-related structured data in the analyzed page, including price, availability, ratings, and product tags.

Canonical: https://docs.equalai.io/ai-visibility/shopping-feed
Markdown: https://docs.equalai.io/ai-visibility/shopping-feed.md

Despite its name, Shopping Feed currently checks product markup in the project's initial page HTML. It does not upload, generate, validate, or submit a merchant product feed, and it does not audit your entire catalog.

## Check a product page

1. Confirm the project URL. A home page without product markup can return little useful commerce data.
2. Open **GEO Optimization → Check Shopping Feed**.
3. Review **Products**, **Price**, **Availability**, and **Ratings**, then the individual checks.
4. Compare the detected data with the actual published product information.
5. Make corrections in your website and rerun from Optimization to retain a new result.

The analysis reads Product JSON-LD from top-level or `@graph` entries, supported nested offers, Product microdata, and product-related Open Graph signals. It does not execute JavaScript to discover later-injected markup.

## Scoring rules

| Signal | Points |
| --- | --- |
| Product JSON-LD or Product microdata | 30 |
| Price information in supported structured data | 25 |
| Availability information | 15 |
| Aggregate rating signal | 10 |
| Product-related Open Graph tags | 20 |

Price detection recognizes supported offer price or low-price fields. A detected value is a presence signal, not validation of the currency, commercial accuracy, or every required offer property. Ratings can come from aggregate-rating markup; the check does not authenticate reviews.

Microdata can satisfy the Product check without supplying all the fields recognized from JSON-LD. Use [Schema](https://docs.equalai.io/ai-visibility/schema) to inspect parsing and discovered types.

## Act on a finding

- No products: verify that the analyzed URL actually represents products and that markup is in the initial response.
- Missing price or availability: check the markup structure and publish truthful, current values where appropriate.
- Missing ratings: add rating markup only when you have legitimate ratings to represent. Its absence may be correct.
- Missing product tags: decide whether commerce previews are appropriate for this page before adding tags.

Passing these checks does not establish eligibility for an AI shopping answer, acceptance by a merchant platform, or inventory synchronization. Validate platform-specific requirements separately.

## Report lifecycle

**Refresh** reanalyzes the current project URL in the open report. **Export** downloads analysis JSON, not a shopping-feed file. **Share** copies an access-controlled page URL. To preserve a new history entry after changing markup, run **Check Shopping Feed** again from Optimization.


---

# Site Structure

Review the analyzed page's links and headings alongside its root sitemap, breadcrumb, and canonical signals.

Canonical: https://docs.equalai.io/ai-visibility/site-structure
Markdown: https://docs.equalai.io/ai-visibility/site-structure.md

Site Structure inspects one page's initial HTML and the root `/sitemap.xml`. It does not crawl an entire website, measure click depth, or reconstruct a complete site hierarchy. In Optimization, the card is **Structure** and the action is **Check Site Structure**.

## Inspect the report

1. Confirm that the project URL is the page you intend to check.
2. Choose **GEO Optimization → Check Site Structure**.
3. Compare sitemap entries, internal and external link counts, and H1/H2/H3 counts.
4. Review breadcrumb, canonical, and heading checks before changing your templates or navigation.
5. Publish changes and rerun from Optimization for a new stored comparison.

## How the score is calculated

| Signal | Points |
| --- | --- |
| Recognizable sitemap XML | 25 |
| At least five internal links | 20 |
| Exactly one H1 | 10 |
| H1 and H2 headings present | 15 |
| Breadcrumb signal present | 15 |
| Canonical link present | 15 |

The heading-hierarchy check tests for H1 and H2 presence; it does not validate the full sequence of heading levels. Breadcrumb detection uses recognizable markup/text signals, not a test of every breadcrumb destination.

## Interpret counts carefully

Internal links are classified by hostname with a leading `www` normalized. Repeated links count more than once. Counts therefore do not establish how many unique pages are reachable or whether destinations return successful responses.

The sitemap check recognizes a `urlset` or `sitemapindex` and counts location entries in the retrieved file. In a sitemap index, entries refer to child sitemaps, not individual pages. EqualAI does not recursively open those child sitemaps in this check. A sitemap published somewhere other than the root path may not be detected.

## Turn findings into changes

- Check that important links exist in the server-returned HTML and use destinations appropriate for visitors.
- Keep headings descriptive and logically organized; a passing presence check alone does not prove good document structure.
- Verify canonical targets and breadcrumb destinations independently.
- If the sitemap signal is missing, inspect the file location and response before generating a replacement.

Do not add redundant links or misleading breadcrumbs just to reach a threshold. Pair this report with [Crawlability](https://docs.equalai.io/ai-visibility/crawlability) and [Discoverability](https://docs.equalai.io/ai-visibility/discoverability) when investigating missing pages.

## Export and rerun

**Export** downloads the displayed JSON analysis. **Share** copies the report URL without granting public access. **Refresh** changes the open view, not the saved report payload; run the check from Optimization when you need a new history entry.


---

# Traffic Analytics

Explore GA4 traffic attributed to recognized AI-related session sources, compare landing pages, and copy results.

Canonical: https://docs.equalai.io/ai-visibility/traffic-analytics
Markdown: https://docs.equalai.io/ai-visibility/traffic-analytics.md

Traffic Analytics measures visits to your website recorded in Google Analytics 4 (GA4). It answers a different question from prompt tracking: which recognized AI-related sources sent measurable traffic, rather than which models mentioned your brand.

## Before you begin

Open **AI Visibility → Traffic Analytics** in the correct project. The page heading is **AI Traffic Analytics**. A workspace administrator must connect a Google account and choose that project's GA4 property. Follow [Connect Google Analytics](https://docs.equalai.io/integrations/google-analytics) if you see **Connect Analytics** or **Select Property** instead of charts.

GA4 must already collect data for the website. A working connection does not guarantee that GA4 has recorded AI referrals.

![Traffic Analytics showing Connect Analytics rather than charts because Google Analytics is not connected](https://docs.equalai.io/images/guides/traffic.webp)

*EarlySEO before Google Analytics is connected. **1.** Connect Analytics setup state. **2.** Start Google Analytics connection. If you see this screen, follow the connection guide before looking for the dashboard controls below; it is not a zero-traffic result.*

## Read the dashboard

1. Choose **Last 7 days**, **Last 30 days**, or **Last 90 days** under **Date range**. The default is 30 days.
2. Leave **Source** on **All Sources**, or select a returned source to narrow the report.
3. Choose **Sessions**, **Users**, **Conversions**, or **Conv. Rate** for the chart. Use **Day**, **Week**, or **Month** to change its grouping.
4. Compare **Top AI Sources Table** with **Top Landing Pages** to identify sources and entry pages worth investigating.

| Metric | Meaning |
| --- | --- |
| AI Sessions | GA4 sessions attributed to the matching session sources. |
| AI Users | GA4 user counts returned for those sources. Sums across source rows or daily buckets are not a deduplicated count of people across the whole period. |
| Conversions | The GA4 conversions metric for the matching traffic. It can contain multiple conversions per session. |
| Conversion Rate | Conversions divided by sessions, multiplied by 100. Zero sessions produces 0%; this is not a percentage of uniquely converting people. |
| Share | A source's session share, even when another chart metric is selected. |

**Conversion Settings** opens Project Settings. The report currently uses GA4's conversions metric; selecting event names there does not filter this report to those events.

## What counts as AI traffic?

EqualAI matches GA4 session-source text against recognized names, including ChatGPT/OpenAI, Perplexity, Gemini/Bard, Claude/Anthropic, Copilot, You.com, and Phind. This is attribution by source name, not a record of every visit influenced by AI. Visits with missing referrers or different source labels may not appear. Matching a source also does not establish a citation, model training, or crawler visit.

The source request returns up to 50 source/medium rows. Landing-page data returns up to 25 rows. These bounded reports are not an unlimited export of your GA4 property.

## Review and copy landing pages

Use **Sessions**, **Users**, or **Path** to sort **Top Landing Pages**, then move through its 10-row pages. **Export** copies CSV to your clipboard; it does not download a file. Paste it into a spreadsheet or text editor. The CSV contains the returned landing-page rows in the current sort order, not just the visible page, with landing page, sessions, and users.

**Annotations** shows available project annotations alongside the chart; it does not change the measurements.

## Freshness and troubleshooting

- **Data through** identifies the latest date in the returned series. Its age describes this report's data, not a guaranteed GA4 processing delay.
- Reports can be reused from a four-hour cache for the same property, range, and source. Reloading is not a guaranteed live refresh.
- For an empty chart, verify the property and selected source, try a wider range, and check whether GA4 itself has matching source traffic. Do not interpret missing attribution as proof that no AI-influenced visits occurred.
- For permission or connection errors, ask an administrator to reconnect with an account that can access the property.
- Use **Change property** to select another accessible property. Subsequent reports use that property; this does not merge properties.
- **Disconnect** stops this project's GA4 connection. See [disconnect consequences](https://docs.equalai.io/integrations/google-analytics#disconnect-or-change-property) before using it.

For crawler requests rather than human sessions, use [AI Agents](https://docs.equalai.io/ai-visibility/ai-agents).


---

# Ads

Investigate captured in-answer ad placements without confusing observations with bidding or spend data.

Canonical: https://docs.equalai.io/analytics/ads
Markdown: https://docs.equalai.io/analytics/ads.md

Open **Reports → Ads**. This Beta report describes ad placements captured with tracked answers. It does not connect to an advertiser's bidding account and does not report verified spend, bids, impressions, or clicks.

## Set the scope

1. Choose the date range.
2. Use **All filters** for **Tags**, **Topics**, **Formats**, and **Engines**.
3. Review **Ad presence over time** and **Ad share by advertiser**.
4. Search and sort **Advertisers in your market** to investigate specific advertisers.

The advertiser search narrows that table; it does not redefine all summary cards or the exported filtered placement set.

## Understand the metrics

| Metric | Meaning |
| --- | --- |
| Ad coverage | Distinct prompts with a captured ad in the filtered rows divided by the project's tracked prompt count |
| Ads in AI answers | Number of captured placement rows, not unique ads or people |
| Ad share | An advertiser's placement count divided by all filtered placements |
| Times seen | Captured placements for the advertiser |
| Creatives | Distinct normalized ad titles in the captured placements |

The ad-coverage numerator responds to report filters, while the denominator remains the project's prompt count. Multiple placements can come from one answer or one advertiser. A captured payload may include an ad that was not visibly rendered; inspect any rendering indicator before claiming the user saw it.

## Read branded-prompt evidence carefully

**Bidding on your brand** lists advertisers whose placements appear for prompts or captured keywords matching your brand aliases. Despite the section title, this does **not** establish that an advertiser purchased your brand keyword or intentionally targeted your business.

Use **Ads by topic** to locate affected question groups, then inspect the prompt and captured answer. Keep observations separate from assumptions about advertiser strategy.

**Export** downloads placement rows in the report's filter scope. An advertiser table search is not an additional export filter.

## Troubleshooting

- **No ads in the window:** no matching placements were captured; this is not proof that the engine never shows ads.
- **Coverage seems lower than expected:** the denominator includes all tracked prompts, not just prompts represented by filtered ad rows.
- **Counts exceed the number of prompts:** one prompt can have multiple placements and repeated runs.
- **A branded match is surprising:** inspect the actual prompt/keyword and your aliases.

## Related guides

[Responses](https://docs.equalai.io/analytics/responses) · [Prompts](https://docs.equalai.io/analytics/prompts) · [Competitors](https://docs.equalai.io/tracking/competitors)


---

# AI Model Insights

Compare mention, citation, sentiment, visibility, and answer availability across AI models.

Canonical: https://docs.equalai.io/analytics/ai-model-insights
Markdown: https://docs.equalai.io/analytics/ai-model-insights.md

Open **Reports → AI Model Insights** for model-level comparisons. Use the same project and comparable response scope before concluding that one model performs better than another.

## Choose a report tab

| Tab | Use it to investigate |
| --- | --- |
| Mention Rate | How often answers name each tracked brand |
| Citation Rate | How often answers cite each brand's domain |
| Net Sentiment Score | The tone scores associated with captured responses |
| AI Visibility Score | Rank-weighted brand visibility, including zero contributions |
| Position Distribution | Where brands appear in recorded mention positions |
| Date Comparison | Differences between two selected dates |
| AI Overview Results | Whether captured Google AI Overview queries returned overview results |

Set the date range and **All filters**, then select competitors where that control is offered. Filters and brand controls are not interchangeable across tabs. **Date Comparison** has its own date selections; inspect those even when the shared date control looks correct.

![EarlySEO AI Model Insights with the metric selector, comparison filters, and brand-by-model matrix marked](https://docs.equalai.io/images/guides/models.webp)

*EarlySEO example snapshot. **1.** Choose the metric. **2.** Keep filters and competitors consistent. **3.** Compare the same metric across brands and models. These values illustrate the report layout, not expected performance.*

## Read shares and rates correctly

**Your Mention Share by Model** distributes your mentioning answers across models. It is not the same as each model's mention rate, which divides by that model's answer count.

**Your Citation Share by Model** distributes responses citing your domain across models. The citation-rate matrix checks each row brand's domain; it does not count every unrelated citation as a brand citation.

**AI Visibility Score** uses reciprocal-rank contributions: 100 at rank 1, 50 at rank 2, 33 at rank 3, and zero for absent brands or invalid/missing ranks. The average includes all answers in the model's scope.

Matrix averages give each model column equal weight. A model with few answers can therefore affect the average as much as one with many answers.

## Interpret sentiment carefully

The own-brand model summary averages the captured response sentiment across answers. The competitor matrix averages that same response sentiment over answers mentioning each brand. It does not independently analyze the tone directed at each competitor. Use [Reputation](https://docs.equalai.io/analytics/reputation) and original [Responses](https://docs.equalai.io/analytics/responses) when brand-specific evidence matters.

## Investigate date and overview differences

In **Date Comparison**, choose the two dates and brand, then review **What changed** and **Changes By**. Keep the other filters consistent.

**AI Overview Results** measures result availability among captured overview queries, not your brand's mention rate. Time buckets with no queries are skipped rather than counted as failed overviews.

Use the tab's **Export** control when available to retain the displayed comparison.

## Troubleshooting

- **A model column is zero or sparse:** check answer counts, filters, and available ranks.
- **The default period seems old:** this report can anchor presets to the latest tracked day rather than today.
- **A share is high but a rate is low:** share depends on volume across models; rate uses the model's own denominator.
- **Sentiment differs from another report:** verify mention-only versus all-response aggregation.

## Related guides

[Metrics reference](https://docs.equalai.io/analytics/metrics) · [Overview](https://docs.equalai.io/analytics/overview) · [Sentiment](https://docs.equalai.io/analytics/sentiment)


---

# Citations

Compare cited sources by domain, host, or page and inspect the answers that used them.

Canonical: https://docs.equalai.io/analytics/citations
Markdown: https://docs.equalai.io/analytics/citations.md

Open **Citations** to investigate the sources recorded in tracked AI answers. A citation occurrence is not a website visit, backlink, or unique person.

## Narrow the source list

1. Set the date range and **All filters**.
2. Choose **All sources**, **You**, **Competitor**, **Social**, **UGC**, or **Third-party**.
3. Select grouping **By domain**, **By host**, or **By page**.
4. Sort by **Citation share** or another available column, then open a source to inspect its supporting pages and prompts.

Domain grouping combines a site's hosts; host grouping keeps hostnames separate; page grouping focuses on individual URLs. Classification into social, user-generated content, or page types is a useful organizing aid, not a guarantee about editorial quality or ownership.

![EarlySEO Citations with report filters, analysis breakdown choices, and domain, host, or page grouping marked](https://docs.equalai.io/images/guides/citations.webp)

*EarlySEO example snapshot. **1.** Search, date and filters. **2.** Compare source, platform, page type or model. **3.** Group by domain, host or page. The displayed sources and counts reflect this example scope, not website visits.*

## Read the table and charts

| Field | Meaning |
| --- | --- |
| Citation share | This source group's citation occurrences divided by all citation occurrences in the filtered source set |
| Citations | Recorded citation occurrences, including repeated use across answers |
| Prompts | Distinct tracked prompts represented by the group's citations |
| Avg. pos. | Average valid recorded citation position, not organic search rank |
| Page mix | Distribution of classified page types |
| Last seen | Latest recorded occurrence for the displayed source evidence |

Citation share is different from a response-based citation rate. For example, a source with 10 of 50 citation occurrences has a 20% citation share even if those occurrences came from fewer than 50 responses. This is an illustrative example, not customer data.

In **Citation analysis**, split trends by **Source type**, **Platform**, **Page type**, or **Model**. **Hide charts** leaves more room for the source table. Platform and page-type panels retain alternative choices when their own facet is selected, so their distributions need not match a narrowly filtered table exactly.

When a comparison period is active, a source can appear because it was cited previously but not in the primary period. Inspect its period labels before treating it as a current citation.

## Inspect a source

Open its detail view and follow the available page, prompt, and response evidence. Compare your domain with specific competitors rather than assuming every third-party source is a competitor. Use **Export CSV** when you need the source table outside EqualAI.

## Troubleshooting

- **A domain appears more than once:** check whether grouping is by host or page.
- **Share changed without a count increase:** the total number of citations in the denominator may have changed.
- **No citations appear:** check filters and whether the selected answers actually contain source data.
- **A page-type label looks wrong:** verify the underlying URL and content; classification is not a manual editorial review.

## Related guides

[Metrics reference](https://docs.equalai.io/analytics/metrics) · [Responses](https://docs.equalai.io/analytics/responses) · [URL Report](https://docs.equalai.io/analytics/url-report)


---

# Dashboard

Find projects, review workspace usage, and choose the project to investigate.

Canonical: https://docs.equalai.io/analytics/dashboard
Markdown: https://docs.equalai.io/analytics/dashboard.md

Open **Dashboard** from the main navigation. This is the workspace entry point; [Overview](https://docs.equalai.io/analytics/overview) is the detailed analytics view for an individual project.

## Find a project

1. Use the project search to narrow the list.
2. Choose a date range when comparing project performance. Available presets include 7 days, 30 days, 90 days, and 12 months, plus a custom range.
3. Open the project you want to investigate, then confirm its name in the project selector.
4. Continue to **Overview**, **Prompts**, or **Responses** for supporting evidence.

Bookmarks help you return to projects. They are stored in the current browser, so they are not a shared workspace preference and may disappear when browser storage is cleared.

## Read workspace usage

The usage area includes **Prompts**, **Executions**, **Projects**, **Content Tasks**, and **Team Members**. These are different resources: a prompt is a tracked question, while an execution is a run of a question against an AI model or surface. One prompt can produce multiple executions over time.

Use the limits actually displayed for your workspace. Do not infer a plan allowance from another project's report totals. Workspace usage and a project card's date-filtered analytics do not have the same scope.

If you have the required access, **Create a New Project** starts [project onboarding](https://docs.equalai.io/getting-started/onboarding). Creating a project does not itself establish that any AI responses have been collected.

## Troubleshooting

- **A project is missing:** clear the search and check that you are in the correct workspace and have access to the project.
- **The project has no recent results:** widen the date range, then check its [prompts and schedule](https://docs.equalai.io/tracking/schedules).
- **A bookmark is missing on another device:** bookmarks are browser-local.
- **Usage and response totals differ:** compare scope first. Workspace allowance counters are not interchangeable with filtered report totals.

## Related guides

- [Project Overview](https://docs.equalai.io/analytics/overview)
- [Billing and plan](https://docs.equalai.io/account/billing)
- [Team members](https://docs.equalai.io/account/team-members)


---

# Analytics and reports

Choose the right report and follow a result back to its tracked answers.

Canonical: https://docs.equalai.io/analytics
Markdown: https://docs.equalai.io/analytics.md

EqualAI analytics describe answers collected for your tracked prompts. They are not a census of every answer an AI service gives, and do not measure website visits or crawler requests.

## Choose a starting point

| Task | Open |
| --- | --- |
| Switch projects and review workspace usage | [Dashboard](https://docs.equalai.io/analytics/dashboard) |
| Review mentions, visibility, citations, and competitors | [Overview](https://docs.equalai.io/analytics/overview) |
| Manage questions and inspect prompt-level results | [Prompts](https://docs.equalai.io/analytics/prompts) |
| Read the answer behind a metric | [Responses](https://docs.equalai.io/analytics/responses) |
| Investigate a cited domain or page | [Citations](https://docs.equalai.io/analytics/citations) |
| Understand tone around your brand | [Sentiment](https://docs.equalai.io/analytics/sentiment) |
| Compare model performance | [AI Model Insights](https://docs.equalai.io/analytics/ai-model-insights) |
| Assess evidence-backed reputation dimensions | [Reputation](https://docs.equalai.io/analytics/reputation) |
| Inspect captured background searches | [Query Fanouts](https://docs.equalai.io/analytics/query-fanouts) |
| Review captured ad placements | [Ads](https://docs.equalai.io/analytics/ads) |
| Review a saved weekly summary | [Weekly Report](https://docs.equalai.io/analytics/weekly-report) |
| Investigate a URL target | [URL Report](https://docs.equalai.io/analytics/url-report) |
| Open a previously saved artifact | [Report History](https://docs.equalai.io/analytics/report-history) |

## Compare like with like

1. Select the same project.
2. Check the date range and any comparison period.
3. Match model, country, language, topic, and tag filters where offered.
4. Read the metric definition before comparing screens. In particular, **Citation rate** does not mean the same thing everywhere.
5. Open the underlying responses before acting on a change.

A percentage based on two answers is less informative than one supported by many repeated runs. A missing answer, unavailable provider evidence, and a recorded zero are different conditions. Clear filters and check run completion before interpreting an empty panel.

For formulas and known differences between screens, use the [metrics reference](https://docs.equalai.io/analytics/metrics). For data collection, see [schedules](https://docs.equalai.io/tracking/schedules). Website visits and automated requests are covered separately in [Traffic Analytics](https://docs.equalai.io/ai-visibility/traffic-analytics) and [AI Agents](https://docs.equalai.io/ai-visibility/ai-agents).


---

# Metrics reference

Understand EqualAI's actual denominators, rank-weighted visibility, and differences between reports.

Canonical: https://docs.equalai.io/analytics/metrics
Markdown: https://docs.equalai.io/analytics/metrics.md

Always identify the screen, project, date range, and active filters before comparing a metric. A familiar label can describe a different calculation on another screen.

## Mentions and visibility

**Brand mentions** counts responses that name a brand. Repeating the name several times in one answer does not create several mentioning responses.

**Mention rate** is `100 × mentioning responses ÷ responses in scope`.

**AI visibility in Overview and AI Model Insights** is an average rank-weighted score. For a response naming the brand at a valid rank, its contribution is `round(100 ÷ rank)`. Rank 1 contributes 100, rank 2 contributes 50, and rank 3 contributes 33. An absent brand or missing/invalid rank contributes zero. Those zeros remain in the response denominator.

For example, three answers contributing 100, 50, and 0 have an average visibility of 50. This is an illustrative example. It is not a search-engine ranking or a percentage of all AI users who saw the brand.

**AI Visibility Reports** uses a different calculation: its visibility percentage is the rounded percentage of evaluated responses mentioning the brand. Do not equate this with the rank-weighted Overview score. See [AI Visibility Reports](https://docs.equalai.io/ai-visibility/ai-visibility-reports).

## Share of voice

Overview share of voice is `100 × a brand's mentioning-response count ÷ the sum of those counts for the project and tracked competitors`.

One response can name several tracked brands and contribute to several brand counts. This is a comparison within the tracked set, not total industry market share. Changing the competitor set can change the denominator.

A change from 20% to 25% is **+5 percentage points**, not +5% relative growth. Share-of-voice movement uses percentage points.

## Citations: check the screen

| Screen or metric | Numerator | Denominator |
| --- | --- | --- |
| Overview citation rate | Responses citing the project domain | Responses in scope |
| AI Model Insights citation-rate matrix | Responses citing the row brand's domain | Responses for that model in scope |
| Responses citation rate | Responses containing any source citation | Responses in scope |
| Citations: Citation share | Citation occurrences for a source group | All citation occurrences in the filtered source set |
| URL Report citation rate | Loaded project responses matching the report target | All loaded project responses |

**Citations** can count multiple occurrences in one response. A citation rate counts a qualifying response once even when it has several qualifying citations.

Overview's project-domain matching normalizes the domain and includes matching subdomains. Do not assume every report applies domain settings identically; investigate the actual source URLs when exact subdomain isolation matters.

## Sentiment

The Sentiment report assigns positive and very positive responses +1, negative and very negative responses −1, and neutral/unknown responses 0. Its net score is the rounded sum divided by **all responses in scope**, multiplied by 100. Unknown responses remain in that denominator; zero is not proof that every answer was classified neutral.

In AI Model Insights, the brand comparison sentiment matrix averages response sentiment among answers mentioning each row's brand. Competitor rows reuse the captured response sentiment; they are not independent competitor-specific tone evaluations. This differs from the own-brand model summary's all-response denominator.

## Aggregation and missing evidence

- AI Model Insights matrix averages weight model columns equally; they are not necessarily pooled response averages.
- Some Overview brand charts carry the last measured value into empty time buckets. This does not create new observations.
- A model with little data can change sharply after a single answer.
- An unavailable rank, no captured citation evidence, and a verified absence are distinct underlying conditions, even when a summary renders zero.
- Ads, reputation dimensions, weekly summaries, and technical checks have their own scope. Follow their guides rather than substituting these formulas.

## Related guides

[Overview](https://docs.equalai.io/analytics/overview) · [AI Model Insights](https://docs.equalai.io/analytics/ai-model-insights) · [Sentiment](https://docs.equalai.io/analytics/sentiment)


---

# Project Overview

Read your project's mention, visibility, citation, and competitor summaries.

Canonical: https://docs.equalai.io/analytics/overview
Markdown: https://docs.equalai.io/analytics/overview.md

Open **Overview** after selecting a project. Use it to identify a change, then investigate the prompts and responses behind that change.

## Set the comparison scope

1. Choose the date range.
2. Use the available filters for tags, models, countries, languages, and competitors.
3. If comparing periods, check the comparison label rather than assuming a calendar-month baseline.
4. Use **Clear all** to remove active filters when the view becomes unexpectedly narrow.

![EarlySEO Project Overview with the reporting period, model and locale filters, and summary metrics marked](https://docs.equalai.io/images/guides/overview.webp)

*EarlySEO example snapshot. **1.** Reporting period. **2.** Model, country and language filters. **3.** Summary metrics. Your results depend on the selected project, period, and filters.*

**Add Prompts** offers several ways to expand tracking, including a single prompt, bulk import, PAA questions, GSC queries, topics, translation, Reddit insights, and reputational prompts. See [Prompt Explorer](https://docs.equalai.io/ai-visibility/prompt-explorer) before saving generated candidates.

## Read the summary cards

| Card | Meaning in Overview |
| --- | --- |
| Responses | Tracked responses in the selected scope |
| Brand mentions | Responses that name your brand, not the number of times its name is repeated |
| Mention rate | Percentage of responses that name your brand |
| AI visibility | Average rank-weighted score; rank 1 contributes 100, rank 2 contributes 50, and absence or an unavailable valid rank contributes zero |
| Citations | Citation occurrences matching your project domain |
| Citation rate | Percentage of responses with at least one project-domain citation |

See the [metrics reference](https://docs.equalai.io/analytics/metrics) for rounding and cross-report differences.

## Investigate a change

- **Mentions:** compare mention counts, mention rate, and AI visibility. More mentions do not necessarily mean stronger positions.
- **Share of Voice:** compare mentions of your project and tracked competitors. This is not market share across an entire industry. Movement is expressed in percentage points against the comparison period.
- **Citations:** identify sources that appear in tracked answers. A source citation and a brand-name mention are different signals.
- **Topic Performance:** find groups of questions that deserve closer investigation.
- **Performance matrix:** compare Visibility, Mention rate, Citation rate, Sentiment, or Share of voice across brands and AI models. **Switch axis** changes the layout; **Export matrix as CSV** exports the matrix view.

Charts use time buckets appropriate to the range. Some brand series carry the last measured value into empty buckets; a flat segment is not proof of a new answer in every interval.

## Troubleshooting

- **No prompts:** add relevant questions first.
- **Prompts exist but no answers:** check whether a run has finished and whether scheduling is enabled.
- **Nothing matches:** clear filters or widen the range.
- **Another screen shows a different rate:** verify its denominator. Responses counts any-source citations, while Overview counts your project-domain citations.

## Related guides

[Prompts](https://docs.equalai.io/analytics/prompts) · [Responses](https://docs.equalai.io/analytics/responses) · [Competitors](https://docs.equalai.io/tracking/competitors)


---

# Prompts

Add and organize tracked questions, review suggestions, and investigate prompt-level evidence.

Canonical: https://docs.equalai.io/analytics/prompts
Markdown: https://docs.equalai.io/analytics/prompts.md

Open **Prompts** for the selected project. A prompt is a question you choose to track; its results are the responses collected for its configured models and runs.

## Add questions to track

1. Select **Add Prompts**.
2. Choose the appropriate creation method. **Single prompt** is useful for a known question; **Bulk import** is useful for an existing list. Other methods help discover candidates.
3. Review the questions, topic, model, country, language, and other settings offered by that method.
4. Save the prompts, then use **Run prompts** to open the run and schedule controls when you are ready to collect answers.

![Add Prompts menu showing the Single prompt and Bulk import choices](https://docs.equalai.io/images/guides/add-prompts.webp)

***1.** Add Prompts menu. **2.** Single prompt or Bulk import. Choose the method that fits your questions; opening this menu does not add or run a prompt.*

Saving a prompt or accepting a suggestion is not the same as completing an execution. See [Prompt Explorer](https://docs.equalai.io/ai-visibility/prompt-explorer) for creation methods and [schedules](https://docs.equalai.io/tracking/schedules) for running questions. Run actions can consume the workspace's execution allowance.

## Organize the tracked list

Use the search box, date range, **All filters**, and topic rail to narrow the list. Columns include **Prompt**, **Type**, **Models**, **Country**, **Responses**, **Mentions**, **Mention Rate**, **Citations**, **Tags**, **Last Update**, and **Creation Date**.

Use tags to compare a consistent subset, such as a product line or campaign. Editors can change prompt details and tags through the available row controls. Viewers can inspect results but cannot mutate prompts.

![EarlySEO tracked prompts with search and filters, topic selection, and prompt rows marked](https://docs.equalai.io/images/guides/prompts.webp)

*EarlySEO example snapshot. **1.** Search, date and filters. **2.** Topic selection. **3.** Tracked prompt rows. The questions and counts shown are examples, not a required tracking set.*

**Export CSV** downloads the table export; the copy control copies CSV to the clipboard. Check the current filters before using an export in a comparison.

> Deleting tracked prompts also removes their tracked responses. Export any results you need before deleting, and review the confirmation carefully. Bulk deletion affects every selected prompt.

## Review suggestions

Open **Suggested** to review project-specific candidates. **Accept** adds a candidate to tracking; **Reject** removes it from consideration. Review wording before accepting: a suggested question is not evidence of search demand or a guarantee that every provider supports the same answer features.

## Inspect a prompt

Open a tracked prompt to inspect its results. Depending on available evidence, the detail view includes **Overview**, **Source distribution**, **Answer signals**, **Brands named by model**, **Read/not cited**, **Query fanouts**, and **AI Chats**.

Use these sections to separate brand mentions, linked citations, and pages read without a citation. Open the underlying response when a result needs explanation. Provider-specific evidence may be absent without indicating a tracking failure.

## Troubleshooting

- **No responses after saving:** wait for a completed run; saving alone does not execute prompts.
- **Actions are unavailable:** confirm editor access and available allowance.
- **A suggestion is missing:** check the correct project and Suggested tab.
- **Counts differ from Overview:** compare filters and scope, then inspect individual responses rather than assuming all cards aggregate identically.

## Related guides

[Responses](https://docs.equalai.io/analytics/responses) · [Tags](https://docs.equalai.io/tracking/tags) · [Metrics reference](https://docs.equalai.io/analytics/metrics)


---

# Query Fanouts

Inspect the background search queries captured while AI engines answered tracked prompts.

Canonical: https://docs.equalai.io/analytics/query-fanouts
Markdown: https://docs.equalai.io/analytics/query-fanouts.md

Open **Reports → Query Fanouts** to inspect captured background searches. A fanout is a query an AI engine used while answering a tracked prompt. It is not necessarily the original prompt, a user search-volume estimate, or a complete retrieval log.

## Find a query

1. Set the date range.
2. Use **All filters** for **Tags**, **Models**, or **Topics**.
3. Search for text in the query, original prompt, topic, model, or tags.
4. In **All queries**, expand the topic and prompt groups to inspect captured queries and the models associated with them.

Repeated normalized query text is grouped together. Occurrence counts and distinct-query counts answer different questions: a query repeated across runs contributes several occurrences but may remain one distinct query.

## Read the summaries

**Brands** counts fanout rows containing your brand or tracked competitor aliases. This is text matching within captured search queries, not evidence that the resulting answer mentioned or recommended the brand. Short or ambiguous aliases can produce matches that need review.

**Common phrases** highlights frequent words and adjacent two-word phrases from the captured query text, after normalization and removal of common filler. These are not independently measured search demand, user intent, or keyword difficulty scores.

Use a recurring phrase as an investigation lead: open the original [response](https://docs.equalai.io/analytics/responses), inspect its cited sources, then decide whether your tracked questions cover that information need.

## Export

**Export** downloads the filtered rows with topic, original prompt, query fanout, model, provider, surface, position, and capture date. The position is captured query metadata, not a guaranteed organic search rank.

## Troubleshooting

- **No fanouts despite having responses:** some providers do not supply background-query evidence. Absence of captured fanouts does not prove no search occurred.
- **Counts fall after filtering:** the summaries are recalculated from matching rows.
- **A brand match seems unrelated:** inspect the actual query and configured aliases.
- **Common phrases look generic:** review the underlying queries rather than treating the phrase summary as a content recommendation.

## Related guides

[Responses](https://docs.equalai.io/analytics/responses) · [Citations](https://docs.equalai.io/analytics/citations) · [Prompt Explorer](https://docs.equalai.io/ai-visibility/prompt-explorer)


---

# Report History

Open and download saved report artifacts for the active project.

Canonical: https://docs.equalai.io/analytics/report-history
Markdown: https://docs.equalai.io/analytics/report-history.md

Open **Reports → Report History** to inspect saved outputs for the active project. The list is scoped to the selected project, not the entire workspace.

## Open an artifact

1. Confirm the project selector.
2. Review **Archive snapshot** for saved-output counts and the latest item.
3. In **Report library**, find the report by **Name**, **Kind**, or **Generated At**.
4. Select **View** to open its detail.

The detail shows report identity and the saved payload. **Payload fields** counts top-level saved fields; it is not a quality score or a count of responses. A generic artifact view can be less visual than the report's dedicated feature page.

Use **Download** for the saved artifact. Use **Back to History** to return to the list. Saving an artifact and running a fresh analysis are distinct operations; opening an existing item does not collect new prompt responses.

## Understand Share

The current **Share** action returns link/embed information. It does not create a verified public publishing workflow or grant another person project access. Do not assume the resulting link opens the exact report detail or bypasses sign-in.

For a sensitive report, confirm recipient access and inspect the downloaded content before sharing it through your approved channel. Do not paste private report payloads into public documentation or support tickets without review.

## Troubleshooting

- **A report from another project is absent:** switch to that project; this is not a workspace-wide list.
- **A report detail is unavailable:** return to history and verify that it still exists and that you have access.
- **The payload is empty or technical:** use the feature's own result page when available. The generic detail preserves what was saved rather than generating a new interpretation.
- **Share displays structured link information:** that is the current behavior, not a completed public share dialog.

Weekly summaries and URL targets also have dedicated history views. Use [Weekly Report](https://docs.equalai.io/analytics/weekly-report) or [URL Report](https://docs.equalai.io/analytics/url-report) when those are the artifacts you need.

## Related guides

[Reputation](https://docs.equalai.io/analytics/reputation) · [Optimization](https://docs.equalai.io/ai-visibility/optimization) · [AI Visibility Reports](https://docs.equalai.io/ai-visibility/ai-visibility-reports)


---

# Reputation

Compare evidence-backed reputation dimensions and inspect how each score is supported.

Canonical: https://docs.equalai.io/analytics/reputation
Markdown: https://docs.equalai.io/analytics/reputation.md

Open **Reports → Reputation** to review your brand and tracked competitors against the dimensions configured for the project. Reputation reports use already collected responses; they do not replace running your prompts.

## Prepare and generate a report

1. Collect answers to questions relevant to the attributes you want to investigate.
2. Open **Configure dimensions** and review the dimensions before saving changes.
3. Select **Generate reputation report** when no report is open, or **Regenerate** to create a new report from available evidence.
4. Use **All reports** to return to the saved report library.

Generating or regenerating is a write action. Changing dimensions and generating reports should be deliberate, not a way to refresh a read-only view. A dimension needs relevant prompt or answer evidence; adding its name does not manufacture supporting observations.

## Read Overview

Set the date range and **All filters**, then inspect **Brand comparison**, **AI reputation score**, and **Dimension breakdown**.

- The overall reputation score averages dimensions that have evidence. Unscored dimensions are excluded, not treated as zero.
- Standing is relative to brands with evidence in the current tracked set.
- Mention share compares tracked brand mentions, not the entire market.
- Dimension coverage distinguishes relevant evidence for a specific attribute from a broad overall signal.
- Visibility and citation authority are indexed relative to the strongest tracked brand. Sentiment uses an absolute scale with 50 representing neutral.

A leader-relative score can change when competitor evidence changes, even without a corresponding change in your own responses. Read response counts, evidence tiers, and reasoning alongside the score.

## Inspect model and raw evidence

Open **Model Comparison** to compare dimension scores by model and focus on the brand or dimension offered by the controls.

Open **Raw Data** for rows by brand, model, and dimension. Use its dimension filter and search, and disable **Hide unscored** when you need to see gaps rather than only supported scores. Read the evidence tier and reasoning before interpreting a number as a robust conclusion.

Use **Export CSV** for the relevant report or raw-data view. Keep the selected scope with the export when sharing a comparison.

## Troubleshooting

- **Generate is unavailable:** collect tracked AI responses first.
- **Responses exist but there is nothing to score:** the selected answers may not mention your project or tracked competitors.
- **A dimension has no score:** inspect whether relevant questions and answers support it; unscored is not poor performance.
- **No rows match:** widen the date range or clear filters.
- **Scores differ from Overview:** these are dimension/evidence and competitor-indexed calculations, not the Overview visibility average.

## Related guides

[Prompts](https://docs.equalai.io/analytics/prompts) · [Sentiment](https://docs.equalai.io/analytics/sentiment) · [Competitors](https://docs.equalai.io/tracking/competitors)


---

# Responses

Read captured AI answers and distinguish brand mentions, citations, and provider evidence.

Canonical: https://docs.equalai.io/analytics/responses
Markdown: https://docs.equalai.io/analytics/responses.md

Open **Responses** to inspect the answers behind your metrics. Confirm the selected project before comparing results.

## Find an answer

1. Set the date range and use search or **All filters**.
2. Choose **All**, **Mentions you**, or **Competitors only** to narrow the answer list.
3. Use the topic and prompt rail to focus on one group of questions.
4. Open an answer and inspect its model, capture date, country, language, sentiment, and mention/citation indicators.

![EarlySEO Responses controls for mention views, topic or prompt selection, sorting, and response layout](https://docs.equalai.io/images/guides/responses.webp)

*EarlySEO example snapshot. **1.** Response filters and mention views. **2.** Choose a topic or prompt. **3.** Sort and change the response view. The displayed scope and counts are an example; open an answer to review its evidence.*

**Export CSV** provides answer-level fields, including the prompt, model, sentiment, mention and citation evidence, and capture information. Treat exported answers as potentially sensitive business information.

## Understand the summary

- **Responses** counts answers in scope.
- **Mention rate** is the percentage naming your brand.
- **Avg. position** summarizes valid recorded mention positions; missing positions should not be read as rank zero.
- **Citation rate** is the percentage of answers containing at least one source citation, regardless of who owns that source.
- **Sentiment** summarizes the tone classifications in the selected answers.

The citation definition here differs from [Overview](https://docs.equalai.io/analytics/overview), where citation rate refers to your project domain. An answer can cite a competitor or independent publication without citing your website.

## Read the evidence

Use **What the user saw** or **Open** when offered to inspect the captured presentation. Other controls include **Copy**, **Link**, and **Why this result**.

| Evidence | How to interpret it |
| --- | --- |
| Brands named | Brands detected in the answer, not necessarily linked sources |
| Sources cited | Sources associated with citations in the captured answer |
| Quoted in the answer | Available evidence that content appeared in the answer |
| Named by the model | Available model-provided source or name evidence |
| Read, not cited | A source encountered during retrieval without a corresponding answer citation |
| Query fanouts | Captured background searches used while answering |

Evidence availability varies by model and provider. No fanout or read/not-cited record does not prove the provider performed no retrieval. Read the answer itself before concluding that an automated classification is correct.

## Troubleshooting

- **The list is empty:** clear search, tab, topic, and other filters; confirm a run completed in the selected period.
- **A source is present but your brand is not mentioned:** citations and mentions are independent signals.
- **Evidence controls are absent:** the provider may not have supplied that evidence.
- **A link opens without matching your current view:** check project access and the destination's filters.

## Related guides

[Citations](https://docs.equalai.io/analytics/citations) · [Sentiment](https://docs.equalai.io/analytics/sentiment) · [Query Fanouts](https://docs.equalai.io/analytics/query-fanouts)


---

# Sentiment

Inspect the tone of tracked answers and find the evidence behind positive or negative results.

Canonical: https://docs.equalai.io/analytics/sentiment
Markdown: https://docs.equalai.io/analytics/sentiment.md

Open **Reports → Sentiment** to investigate how captured answers describe your brand. Sentiment is an automated classification of answer tone, not a customer survey or a measure of sales performance.

## Review the overview

1. Set the date range and **All filters**.
2. Start on **Overview** and compare **Sentiment over time**, **Share of sentiment**, and **Sentiment distribution**.
3. Use **Net sentiment by model** to identify model differences.
4. Check **Brand reach** separately: competitor reach shows mentions, not independently scored competitor sentiment.

![EarlySEO Sentiment report header showing its summary mix, report views, and search and date filters](https://docs.equalai.io/images/guides/sentiment.webp)

*EarlySEO example snapshot of the report's top controls. **1.** Summary and sentiment mix. **2.** Switch report views. **3.** Search, date and filters. Use the sections below these controls to investigate the displayed classifications.*

The sentiment categories are **Very Positive**, **Positive**, **Neutral**, **Negative**, **Very Negative**, and **Unknown**. Unknown means no usable classification; do not treat it as evidence of neutral language.

## Understand the net score

Positive categories contribute +1, negative categories −1, and neutral/unknown categories 0. EqualAI divides the total by all responses in scope and multiplies by 100, then rounds. Thus unknown responses reduce the magnitude of the score without adding positive or negative evidence.

A score of +100 means every response in that scope contributed positively. A score of zero can mean neutral/unknown answers or a balance of positive and negative answers. Read the distribution before deciding which explanation applies.

The time chart contains days with responses. A gap does not establish stable sentiment during the missing period.

## Investigate individual answers

- Open **Details** to inspect the prompt, model, answer, classification, and available sentiment explanation.
- Use sentiment, topic, search, and sorting controls to find the answers that need review.
- Open **Topics** to compare question groups.
- Open **By Tag** to compare your own tracking segments.
- Use **Export CSV** for an external review of the report data.

Read the original answer before acting on a negative label. Quoted criticism, comparisons, and mixed-tone answers can require human interpretation.

## Troubleshooting

- **Many answers are Unknown:** inspect response evidence and model coverage instead of assuming neutral perception.
- **A model score is volatile:** compare its response count and period with the other models.
- **Net score and distribution seem inconsistent:** unknown answers remain in the net-score denominator.
- **Competitor sentiment is missing:** Brand reach is a mention comparison, not a competitor tone report.

## Related guides

[Responses](https://docs.equalai.io/analytics/responses) · [Metrics reference](https://docs.equalai.io/analytics/metrics) · [Reputation](https://docs.equalai.io/analytics/reputation)


---

# URL Report

Inspect citations associated with a URL target and understand the current Beta matching limitations.

Canonical: https://docs.equalai.io/analytics/url-report
Markdown: https://docs.equalai.io/analytics/url-report.md

Open **Reports → URL Report**. This Beta workflow saves a target for examining recorded citation evidence; it is not a crawl, live search-ranking check, or traffic report.

## Choose a target

The entry screen offers **Match mode**, **Terms (comma-separated for OR)**, **Only my domain**, and **Run report**. Match-mode choices include **URL (exact)**, **URL contains**, **Path contains**, **Host**, **Host or subdomain**, and **Domain**.

> The current result view does not consistently honor every entry option. Use a single full URL or hostname and verify the supporting evidence before relying on the result. Do not rely on exact-URL, multi-term OR, cross-project, or subdomain-mode isolation in this Beta view.

For a conservative investigation:

1. Confirm the active project.
2. Enter one known cited URL or hostname.
3. Select **Run report** only when you intend to save a report target.
4. Inspect the resulting **Prompt evidence** and compare with [Citations](https://docs.equalai.io/analytics/citations).

The current result matching uses the first saved term, matches its normalized host, and, when a path is present, uses a path prefix. A target ending in `/guide` can therefore include longer paths beginning with `/guide`; the **URL (exact)** label should not be interpreted as strict equality in the result view.

## Read the result

- **AI citations:** matching citation occurrences across citing responses.
- **Citation rate:** responses with a matching citation divided by loaded responses for the active project.
- **Brand mentions:** citing responses that also mention your brand.
- **Brand sentiment:** the most common sentiment label among citing responses; the current summary groups unknown/empty labels with neutral.
- **Prompt evidence:** the prompts, models, dates, positions, and mention/sentiment information supporting the result.

The report does not apply the broader scope implied by every entry setting. Use evidence rows, not a selected mode alone, to establish what was included.

## Saved reports and limitations

Use **View saved URL reports** or **History** to return to saved targets. **Refresh** updates the saved target; **Delete** removes it. Neither should be used during a read-only review.

Suggested URLs currently come from saved targets rather than a verified ranking of the most-cited URLs. The **Owned** badge and “last 90 days” suggestion label should not be used as independent proof of ownership or recency.

## Troubleshooting

- **A path includes more pages than expected:** verify the prefix-matching behavior above.
- **Several terms do not produce the expected union:** investigate each term separately or use Citations filters.
- **No evidence appears:** confirm the active project has captured citations for the target.
- **Sentiment says neutral:** inspect the underlying classifications; unknown entries may be included.

## Related guides

[Citations](https://docs.equalai.io/analytics/citations) · [Responses](https://docs.equalai.io/analytics/responses) · [Metrics reference](https://docs.equalai.io/analytics/metrics)


---

# Weekly Report

Open a saved weekly visibility summary, inspect its period, and download its PDF.

Canonical: https://docs.equalai.io/analytics/weekly-report
Markdown: https://docs.equalai.io/analytics/weekly-report.md

Open **Reports → Weekly Report** for **Your weekly visibility archive**. Weekly reports are saved summaries, not the same as a live dashboard with your current filters.

## Open an existing report

1. Review the latest report's period and score.
2. Select **Open latest**, when available, or choose a report from **The archive**.
3. Check the report's reporting period before comparing it with Overview.
4. Use **PDF** in the report detail to download the report, or return to the archive with its back link.

The archive shows status labels such as **Leading**, **Gaining**, **Steady**, **Slipping**, and **At risk**. Treat these as report summaries, not guarantees about business outcomes.

## Generate deliberately

**Generate this week's report** requests a new report. This is a generation action, not merely a browser refresh. The report needs response evidence for the relevant week. Allow generation to finish before expecting a new archive entry.

Do not assume opening the archive runs prompts. If the week has no evidence, first review the [prompt schedule](https://docs.equalai.io/tracking/schedules) and completed runs. Availability of a saved report does not by itself confirm that an email was delivered.

## Read the report

Depending on available evidence, sections include:

- **The scorecard**, with baseline or week-over-week context.
- **By AI surface** and **Share of voice**.
- **Brand sentiment** and **Where AI looked**.
- **Prompt movers**, including gains, losses, and gaps.
- **This week's wins**, **Active risks**, and **Your plan for next week**.

Use the recorded comparison period and metric units. **New this week** indicates a missing prior comparison rather than a measured improvement from zero. Review supporting prompts and sources before treating a suggested next step as a confirmed diagnosis.

## Troubleshooting

- **The archive is empty:** collect response evidence in the relevant week before requesting a report.
- **Generation has not produced an entry:** allow it to complete, then return to the archive; avoid repeatedly submitting the action.
- **A detail link is unavailable:** the report may have been removed or the link may be stale. Return to the archive.
- **The score differs from a live view:** compare reporting period and saved evidence, not just the label.

## Related guides

[Overview](https://docs.equalai.io/analytics/overview) · [Report History](https://docs.equalai.io/analytics/report-history) · [Schedules](https://docs.equalai.io/tracking/schedules)


---

# Getting started with EqualAI

Set up a project, choose what to track, and understand which settings belong to your project, workspace, or account.

Canonical: https://docs.equalai.io/getting-started
Markdown: https://docs.equalai.io/getting-started.md

EqualAI tracks how AI answers mention a brand and cite websites. Start with a project for the brand or website you want to monitor, then add the prompts and competitors that make the results useful to your team.

## Understand the three scopes

| Scope | What it controls |
| --- | --- |
| Project | Brand identity, target country and language, tracked prompts, competitors, tags, annotations, and automatic prompt runs. |
| Workspace | Subscription, shared project and prompt capacity, AI credits, and workspace owner/admin access. |
| Your account | Your name, password, display currency, and app timezone preferences. |

Selecting a different project changes the context for project pages. It does not create a separate subscription. A workspace admin can access all its projects; an invited member or viewer receives access to the selected project, not every project in the workspace.

## Before you begin

Have the following ready:

- The website domain and recognizable brand name.
- A main country and language for tracking.
- Questions your audience might ask an AI assistant, or topics from which to develop them.
- At least one relevant competitor.
- A decision about which models and run frequency fit your available credits.

You need available project capacity to create another project. Workspace billing changes require a workspace owner or admin. If you joined through an invitation and cannot see a project, check that you accepted the invitation with the intended account.

## Set up your first project

1. Enter your website and targeting preferences.
2. Review the suggested project details and any alternative brand names.
3. Review topics or choose to supply your own prompts.
4. Review competitors and their identifiers.
5. Review or import prompts.
6. Choose initial tracking models and schedule.
7. Review the website report, then the available plan choices.

See [Onboarding](https://docs.equalai.io/getting-started/onboarding) for the exact controls, import limits, and recovery steps. Review suggested content rather than treating it as a verified description of your business.

## Build a useful tracking set

Prefer prompts that represent distinct customer questions. Separate brand-specific questions from broader category questions so you can interpret the results in context. Use [tags](https://docs.equalai.io/tracking/tags) to group prompts by intent, audience, or product area.

**Synthetic example:** a fictional outdoor retailer could track “What should a beginner look for in a hiking backpack?” and tag it `Buying guide`. This is an example of prompt organization, not customer data or a promised result.

Add [competitors](https://docs.equalai.io/tracking/competitors) that genuinely overlap with your audience. Their names, domains, and aliases help identify them in AI answers; they are not additional projects.

## Run and maintain tracking

Open the schedule/run dialog to review model access and the displayed credit estimate before running prompts. A manual run is queued work: submitting it is not the same as receiving completed answers. [Schedules](https://docs.equalai.io/tracking/schedules) explains automatic runs, selected-prompt runs, fan-out, and Instant Run.

Use [annotations](https://docs.equalai.io/tracking/annotations) to record changes such as a content update. An annotation adds context; it does not prove that an event caused a later movement in results.

## When something is unavailable

- **No project:** select an accessible project or ask a workspace admin to invite you.
- **Read-only access:** a viewer cannot make tracking changes. Ask an admin to review your role.
- **Capacity or credit warning:** a workspace admin can review [billing and usage](https://docs.equalai.io/account/billing). Prompt capacity and execution credits are different limits.
- **No results yet:** review the run status and wait for queued work to complete before drawing conclusions.
- **Setup or access error:** use [support](https://docs.equalai.io/account/support) and include the visible error without passwords or tokens.

## Next steps

- [Review project settings](https://docs.equalai.io/account/project-settings).
- [Set personal account preferences](https://docs.equalai.io/account/account-settings).
- [Invite a teammate with the right access](https://docs.equalai.io/account/team-members).
- [Connect an AI assistant with read-only access](https://docs.equalai.io/account/ai-assistant-access).


---

# Create and configure a project

Work through website setup, topics, competitors, prompts, scheduling, and plan activation without confusing project and workspace settings.

Canonical: https://docs.equalai.io/getting-started/onboarding
Markdown: https://docs.equalai.io/getting-started/onboarding.md

Onboarding establishes what EqualAI should recognize and track for a project. Review the generated suggestions before continuing; your website may have several products, regional brands, or similarly named competitors.

## 1. Enter your website

On **Enter your website**, provide the **Domain**, **Main Country**, **Main Language**, and **App timezone**.

- Use the domain you intend to monitor, rather than a competitor's website.
- **Match exact subdomain only** narrows owned-citation matching to that exact host. Without exact matching, a host and its subdomains can count as owned.
- Review the timezone after choosing a country. It is your app display preference, not a second target market.
- **Send me weekly visibility reports for this project** controls the project's email-report preference.

Continue and allow the website analysis to finish. EqualAI uses it to prefill later steps. If project capacity is exhausted or subscription access blocks setup, review the workspace subscription rather than repeatedly creating drafts.

## 2. Review Project Details

Check **Name**, **Industry**, and **Description**. Make the description specific enough to distinguish this brand from unrelated organizations.

Use **Alternative names** for distinctive former names, brand spellings, or acronyms that AI answers may use. The project name and domain are already matched; do not add the domain again.

Up to 12 additional names are accepted. Each should be 2–60 characters and no more than six words. Very short names must be uppercase acronyms; common fragments and redundant variants may be rejected. Follow the validation message rather than adding broad terms that could match unrelated brands.

## 3. Review Topics

Edit or remove suggested topics and use **Add custom topic** to add a missing subject, within the limit shown on screen.

Choose **Have your own prompts? Skip suggestions →** if you already have questions to track. Changing topics can clear generated prompts, so finish the topic review before carefully curating generated questions.

## 4. Review Competitors

Review the proposed **Competitors**, including their domains and identifiers. Discard irrelevant suggestions and add relevant alternatives. Include at least one competitor to complete setup.

Aliases can help recognize abbreviations, product names, or common misspellings associated with that competitor. Avoid generic identifiers that could describe many companies. Competitor settings belong to this project and are subject to its competitor limit.

## 5. Review prompts

In **Review prompts**, inspect the questions grouped by topic. Remove questions that do not match your tracking goal. At least one prompt is needed.

To add your own questions, use **One prompt** or **Paste list**. For a pasted list:

1. Put one prompt on each line.
2. Keep each prompt between 3 and 500 characters.
3. Paste no more than 500 candidates at once, and stay within available prompt capacity.
4. Review the accepted and held-back counts before continuing.

Leading bullets and numbered-list prefixes are removed. Equivalent entries differing only in case, spacing, or trailing punctuation may be treated as duplicates. Imported prompts are organized into topics without generating extra prompts.

**Synthetic example:** `Which features matter when choosing a hiking backpack?` is a sample question for a fictional retailer, not a customer prompt.

## 6. Choose initial tracking

Review the models and weekly run days in the onboarding schedule. Model availability and credit costs can differ. Choose a cadence that leaves room for manual investigations as well as routine tracking.

After setup, the full [schedule dialog](https://docs.equalai.io/tracking/schedules) provides weekly or monthly configuration and explicit Standard or Instant manual runs. Check its upcoming-run preview before relying on a time.

## 7. Review the report and plan

The website report offers a starting assessment. Use **Try again** if the report cannot be built, or **Continue** to proceed. Report generation and email delivery may finish separately; a pending email is not necessarily a failed report.

The plan step shows the available monthly and annual options. Review the current price, capacity, billing interval, and checkout confirmation. Existing subscription access can change which activation steps are needed. Do not assume that creating a project starts a trial or gives it a separate subscription.

## Complete team setup

If post-payment setup appears, answer the role, function, goal, and referral questions. These describe how you intend to use EqualAI; the invitation role is the control that grants teammates access.

- **Company Admin:** workspace-wide access, including billing and team management.
- **Company Member:** editing access to the selected project.
- **Company Viewer:** read-only access to the selected project.

Add each email and choose **Send invite**, or **Skip for now**. Invitations can be managed later in [Team Members](https://docs.equalai.io/account/team-members). **Explore insights** opens the application when setup is complete.

## Resume or troubleshoot

A saved draft can resume at an incomplete step. If returning to an earlier step changes suggestions, review downstream prompts and competitors again. Do not use a new-project draft to repair an existing project: use [Project settings](https://docs.equalai.io/account/project-settings).

If a save fails, keep the visible error, confirm the selected project and your permissions, and check [billing](https://docs.equalai.io/account/billing) for capacity restrictions. For persistent setup failures, contact [support](https://docs.equalai.io/account/support).


---

# EqualAI documentation

Set up your project, measure how AI answers represent your brand, and turn reports into a focused investigation.

Canonical: https://docs.equalai.io/
Markdown: https://docs.equalai.io/index.md

EqualAI tracks answers to your selected prompts so you can understand when your brand appears, which sources AI models cite, and how those results compare with tracked competitors. These guides follow the names and controls in the application.

## Start with a project

1. [Create and configure your project](https://docs.equalai.io/getting-started/onboarding).
2. [Add prompts](https://docs.equalai.io/analytics/prompts) that reflect questions your audience asks.
3. [Add competitors](https://docs.equalai.io/tracking/competitors) and [organize prompts with tags](https://docs.equalai.io/tracking/tags).
4. Read [Project Overview](https://docs.equalai.io/analytics/overview), then open the underlying [responses](https://docs.equalai.io/analytics/responses).

A scheduled or manually requested run needs to finish before new results appear. An empty report does not establish that your brand is absent from AI answers everywhere.

## Understand your results

- **[Dashboard](https://docs.equalai.io/analytics/dashboard):** navigate projects and review workspace usage.
- **[Metrics reference](https://docs.equalai.io/analytics/metrics):** understand the difference between visibility, mention rate, share of voice, and citation rate.
- **[Citations](https://docs.equalai.io/analytics/citations):** investigate the domains and pages used as sources.
- **[Reports](https://docs.equalai.io/analytics):** explore sentiment, model differences, reputation, and changes over time.

## Investigate your website

Use [Traffic Analytics](https://docs.equalai.io/ai-visibility/traffic-analytics) to investigate connected analytics data, [AI Agents](https://docs.equalai.io/ai-visibility/ai-agents) for crawler activity, and [Optimization](https://docs.equalai.io/ai-visibility/optimization) to inspect site checks. These measure different things from tracked-prompt visibility; do not substitute one for another.

## Manage your workspace

Find [project settings](https://docs.equalai.io/account/project-settings), [team access](https://docs.equalai.io/account/team-members), [billing](https://docs.equalai.io/account/billing), and [integrations](https://docs.equalai.io/integrations). Some controls require a project or workspace administrator. The relevant guide explains access restrictions.

## Find the right explanation

Search by a screen name, metric, or task. Each article includes troubleshooting and related guides. Keep the same project, date range, model, and other filters when comparing results across screens.

[Open EqualAI](https://app.equalai.io) to follow along. For an AI assistant, the same article corpus is available as [plain text](https://docs.equalai.io/llms-full.txt), with a [documentation index](https://docs.equalai.io/llms.txt).


---

# Cloudflare

Connect a scoped Cloudflare API token, select a zone, and maintain the AI Agents analytics connection.

Canonical: https://docs.equalai.io/integrations/cloudflare
Markdown: https://docs.equalai.io/integrations/cloudflare.md

Cloudflare supplies request analytics to [AI Agents](https://docs.equalai.io/ai-visibility/ai-agents). You need the correct zone and permissions, not merely a website with the same name as your EqualAI project. Available analytics also depends on Cloudflare access and your account's capabilities.

## Prerequisites

- Workspace administrator access in EqualAI to connect, select, sync, or disconnect the integration.
- Access to the intended Cloudflare zone.
- A Cloudflare API token with **Zone:Read** and **Analytics:Read** for the zones you intend to inspect.
- Any applicable EqualAI plan access for the feature.

Use a narrowly scoped token. Do not use a global API key or include unrelated zones simply for convenience.

## Connect a zone

1. Select the correct EqualAI project and open **AI Visibility → AI Agents**.
2. In **1. Connect Cloudflare**, choose **Create token**. **Copy link** provides the same token-creation destination if needed.
3. In Cloudflare, review the prefilled permissions, choose the appropriate zone scope, and create the token.
4. Paste it into **Cloudflare API Token** and choose **Connect Cloudflare**.
5. In **2. Select zone**, locate the intended zone and choose **Use this zone**.
6. In **3. Analytics**, confirm **Connection**, **Zone**, **Data**, and **Updated**, then select **7D**, **30D**, or **90D**.

![Agent Analytics setup with connection stages, the scoped Cloudflare token creation entry, and the local log preview option marked](https://docs.equalai.io/images/guides/agents.webp)

*EarlySEO before a Cloudflare connection is configured; the token field is empty. **1.** Connection, zone selection and analytics. **2.** Create a limited Cloudflare API token. **3.** Optional local log preview. The preview is separate from connecting a zone.*

The zone list retrieves up to the first 50 accessible zones. If an intended zone is missing, check token scope and permissions; narrowing the token to the intended zone can avoid an oversized selection list.

## Refresh and disconnect

The zone-selection screen offers **Refresh Cloudflare zones** and **Sync now**. Use them after permission or scope changes. If the UI requires a token again, paste an authorized token; you can also supply a replacement. The connection can retain a stored credential beyond its short-lived session cache, so do not interpret a cache message as a security guarantee that all token access has expired.

On the dashboard, **Sync** refreshes connection data and requests updated analytics. Analytics can otherwise be cached for 30 minutes. A failed refresh can leave older data visible; check the error and timestamp rather than assuming the chart is current.

**Disconnect** clears EqualAI's stored Cloudflare credential and marks the connection disconnected. Existing integration metadata and saved analytics are not a guaranteed erasure target. It does not delete Cloudflare logs or revoke the token at Cloudflare.

If the token should no longer be usable anywhere, revoke it in Cloudflare as a separate action. Coordinate first if that token is shared with another authorized application. Reconnect with a new scoped token when needed.

## Troubleshooting

- **No zones attached:** verify the token's zone resources and permissions, then use **Sync now**.
- **Token or permission error:** create or supply a token with the required permissions for the target zone, then reconnect or sync.
- **No request data:** verify the zone and selected period and distinguish a successful zero response from a displayed Cloudflare error.
- **Total requests but no recognized bots:** user-agent matching and returned-group limits can omit activity. Read [AI Agents scope](https://docs.equalai.io/ai-visibility/ai-agents#understand-scope-and-freshness).

For a separate browser-only preview without connecting Cloudflare, use **Analyze server logs** on the connection screen. It is not an upload or a replacement for a zone connection.


---

# Google Analytics

Connect a GA4 property to the current project, verify traffic data, and manage property changes or disconnection.

Canonical: https://docs.equalai.io/integrations/google-analytics
Markdown: https://docs.equalai.io/integrations/google-analytics.md

Connect Google Analytics 4 to use [Traffic Analytics](https://docs.equalai.io/ai-visibility/traffic-analytics). EqualAI reads analytics data; this setup does not install a tracking tag, configure collection on your website, or import every event from the property.

## Prerequisites

- Workspace administrator access in EqualAI for connection, property selection, and disconnection.
- A Google account with access to the intended GA4 property.
- A GA4 property already collecting your website's traffic.
- Google OAuth configured for your EqualAI deployment. A disabled **Google OAuth not configured** control requires deployment administration, not a different property selection.

## Connect and select a property

1. Select the correct EqualAI project and open **AI Visibility → Traffic Analytics**.
2. On **Connect Analytics**, choose **Connect with Google**. If offered, **Reconnect with Google** starts a new authorization flow.
3. Complete Google's account and permission screens using the account that can access your property.
4. On **Select Property**, review **Choose your property** and select the desired property row.
5. Confirm the selected property in the dashboard, then choose a date range and review the data status.

The selected property belongs to this EqualAI project's configuration. Connecting an account alone is not a substitute for selecting the property.

![Connect Analytics screen with the disconnected setup state and Google Analytics connection entry marked](https://docs.equalai.io/images/guides/traffic.webp)

*EarlySEO with Google Analytics not yet connected. **1.** Connect Analytics setup state. **2.** Start Google Analytics connection. This is the connection entry point, not a selected property or a traffic report.*

## Verify the result

Use **All Sources** and a suitable date range first. Compare the matching source traffic with GA4 before assuming an empty result is a connection problem. The report only includes recognized AI-related session sources, not all property traffic.

The same property/range/source report can be reused for four hours. **Data through** reflects the latest returned series date; neither reloading the page nor a recent connection guarantees today's complete data.

**Conversion Settings** opens project settings, but the current report reads GA4's conversions metric rather than filtering to the event-name list stored there. See [metric definitions and export behavior](https://docs.equalai.io/ai-visibility/traffic-analytics).

## Disconnect or change property

Use **Change property** to choose another accessible GA4 property. Subsequent reports use that property; the operation does not combine properties or modify GA4's underlying records.

Use **Disconnect** to remove the stored GA4 credential, mark the integration disconnected, and clear the project's selected GA4 property. Reconnect and select a property before expecting new authenticated reports. This is not a report-history erasure operation and does not delete analytics in Google.

EqualAI's disconnect is not a Google-side consent revocation. If access must also be revoked at Google, review the application's access in your Google account. Revoking a Google grant can affect other features using that account authorization, so coordinate with your administrator.

## Troubleshooting

| Symptom | Next check |
| --- | --- |
| No GA4 properties found | Confirm the Google account's property access, then reconnect with the appropriate account. |
| Permission or authorization error | Ask a workspace administrator to reconnect and review Google's requested consent. |
| Connected but zero AI traffic | Check the property, date range, and GA4 session-source attribution. Missing AI referrers are not proof of no AI influence. |
| Old data | Review **Data through** and allow for the report cache and upstream availability. |
| Export did not download a file | **Export** in Top Landing Pages copies CSV to the clipboard. Paste it into a spreadsheet. |

Do not share OAuth tokens or private analytics exports while troubleshooting.


---

# Google Search Console

Connect Search Console inside Prompt Explorer and turn available search-query rows into reviewed prompt drafts.

Canonical: https://docs.equalai.io/integrations/google-search-console
Markdown: https://docs.equalai.io/integrations/google-search-console.md

Google Search Console supplies search queries to [Prompt Explorer](https://docs.equalai.io/ai-visibility/prompt-explorer). This connection helps you choose questions to track; it does not measure AI referrals, execute prompts, or import Google's historical results as AI responses.

## Prerequisites

- A workspace administrator must authorize the Google connection and change its selected site.
- The Google account must have access to a suitable Search Console property.
- The EqualAI project should have the intended website domain configured so its site can be matched correctly.
- Google OAuth must be available in the EqualAI deployment.
- Creating the resulting prompts requires project editing permission and available prompt capacity.

No separate GA4 connection is required for this research method. Search Console and GA4 are distinct data sources.

## Connect and use queries

1. Select the correct project and open **AI Visibility → Prompt Explorer**. The page heading is **Prompt Research**.
2. Find **Google Search Console** and choose **Connect & start** when disconnected, or start the method and use **Connect Google** when offered.
3. Complete Google's account and permission screens with an authorized account.
4. Return to the workflow, confirm the target country/language, and continue to **Configure source**.
5. Confirm **Search Console site**. Use the selector if a choice is available; the query view is restricted to the configured site rather than an unrestricted account-wide browser.
6. Select relevant query rows and a **Prompt style**: Question, Comparison, or Buyer.
7. Choose **Generate drafts**, review the wording and topics, then choose **Create N prompts**.

If reconnection is needed, the workflow can show **Reconnect Google**. You may need an administrator to complete it.

## Understand the query sample

The request covers a date range starting 28 days before the request date and ending on the request date, returning up to 25 query rows. This is a bounded sample, not every query in the property. Upstream reporting delays or privacy filtering can affect availability.

Rows include query text, clicks, impressions, and position. The selected-row average position shown in the workflow is an arithmetic average, not an impression-weighted position for the entire site.

Selecting a target country or language sets prompt-workflow context. It does not add country or language filters to this Search Console data request. Drafts based on a query may be rewritten into a question or comparison rather than copied verbatim.

## Connection removal and saved prompts

The current Prompt Explorer wizard does not offer a dedicated Search Console **Disconnect** button. Do not use Traffic Analytics' disconnect as a substitute for a Search Console-specific control.

For provider-side access removal, an administrator can review and revoke the application's grant in the connected Google account. This prevents future authorized reads once credentials can no longer be used, but does not delete Search Console data or prompts already created in EqualAI. Revocation may affect other features sharing that Google authorization, and a displayed connection status may not change until an access attempt detects the problem.

To restore access, use **Reconnect Google** with the appropriate account and verify the displayed site again.

## Troubleshooting

- **No queries:** check that the configured property is correct and has query data for the requested period.
- **No configured site:** review the project domain and account access, then reconnect as an administrator.
- **Permission or expired authorization error:** reconnect; do not repeatedly generate drafts from an empty table.
- **Generate drafts unavailable:** select at least one query first.
- **Unexpected market:** remember that target-market selection does not filter Search Console rows. Review and edit the resulting prompts before creation.


---

# Integrations

Connect the data source needed by a visible EqualAI workflow and understand permission and disconnection boundaries.

Canonical: https://docs.equalai.io/integrations
Markdown: https://docs.equalai.io/integrations.md

EqualAI exposes connection setup inside the feature that uses it. Start with the workflow below rather than assuming every provider mentioned in a connection screen is an available integration.

| Connection | Open in EqualAI | What it supplies |
| --- | --- | --- |
| [Google Analytics 4](https://docs.equalai.io/integrations/google-analytics) | AI Visibility → Traffic Analytics | Sessions, users, conversions, and landing pages attributed to recognized AI-related sources. |
| [Cloudflare](https://docs.equalai.io/integrations/cloudflare) | AI Visibility → AI Agents | Zone request totals and recognized crawler user-agent activity. |
| [Google Search Console](https://docs.equalai.io/integrations/google-search-console) | AI Visibility → Prompt Explorer → Google Search Console | Search queries to turn into draft tracked prompts. |
| [AI Assistant Access](https://docs.equalai.io/account/ai-assistant-access) | AI Assistant Access in the account/project menu | Approved assistant access and the available connection-management workflow. |

These connections serve different purposes. Search Console queries do not become GA4 sessions, Cloudflare requests do not establish brand mentions, and connecting a provider does not automatically run tracked prompts.

## Prepare a connection

1. Select the intended EqualAI project.
2. Ask a workspace administrator to connect or change analytics/search credentials.
3. Use a provider account with access to the correct property or zone, and grant only the required access.
4. Complete the feature's selection step and confirm the displayed property, site, or zone before interpreting results.
5. Check the status, data time, and any error. Connected does not necessarily mean data is available for the selected range.

OAuth or service configuration may be unavailable in a particular deployment, and plan or permission restrictions can block a connection. Follow the message shown rather than treating a disabled button as a successful setup.

## Disconnect deliberately

Disconnecting in EqualAI and revoking access at the provider are separate operations. A local disconnect removes or disables the saved connection; it does not delete provider data or necessarily revoke Google consent or a Cloudflare token. It also should not be treated as an erasure of previously saved reports or project prompts.

Read the relevant guide for the controls actually offered. In particular, the Search Console prompt wizard does not currently expose a dedicated disconnect button. Never paste a token into prompts, logs, screenshots, or support messages.

Other analytics provider cards that offer a request action are not working analytics connections. Use the documented available connection rather than assuming a request immediately enables a provider.


---

# Record annotations

Add project-shared notes for reporting context, filter by category, and understand annotation dates and shared tag categories.

Canonical: https://docs.equalai.io/tracking/annotations
Markdown: https://docs.equalai.io/tracking/annotations.md

Annotations record events that help a team interpret results, such as a page update or the start of a campaign. They provide context, not evidence that the recorded event caused a change in visibility.

Annotations on this page are shared with the project. Do not put private notes, credentials, or personal information into a title or description.

## Create an annotation

Select the intended project and open **Annotations**. Creating annotations requires project editing access and a plan that includes the feature.

1. Choose **New Annotation**.
2. Enter a descriptive **Title**.
3. Add a **Description (optional)** explaining what changed.
4. Choose a **Category**, or leave **No category**.
5. Choose a **Color**.
6. Choose **Create Annotation**.

![Empty EarlySEO Annotations page with Annotations and Categories tabs, the create action, and category and search filters marked](https://docs.equalai.io/images/guides/annotations.webp)

*This EarlySEO example has no saved annotations. **1.** Annotations and Categories. **2.** Create an annotation. **3.** Filter by category or search. Start with the create action to record your own event.*

The new entry appears in the table with its date, title, category, description, type, and color. **Type: User** identifies a user-created annotation; it is not a privacy setting.

### Date behavior

Annotations use their creation date. Although the creation form includes a **Date** field, selecting an earlier date does not backdate the annotation. If you are documenting an older event, put its actual date in the description so teammates can distinguish the event from the day the note was added.

**Synthetic example:** `Updated backpack buying guide` could have a description such as `Example event: guide revised on 10 June; note recorded later.` This is illustrative content, not a customer event.

Do not assume annotations appear on every chart. Use the reporting views that support annotation markers, and check their date range when looking for a note.

## Find, edit, or delete a note

Use the category filters and search to narrow the table. Search matches title, description, and category. **No annotations match your filters** can mean that your filter is too narrow, not that the project has no annotations.

Use the row's pencil action to open **Edit Annotation**. Review the **Title**, **Description**, **Category**, and **Color**, then save. Changing those fields does not change the creation date.

Use the delete action and confirm **Delete Annotation** to remove the note. Deletion is a project change; confirm that the team no longer needs its context.

## Manage categories

The **Categories** tab uses the same tags as [Tags](https://docs.equalai.io/tracking/tags). A category created here becomes available for prompt tagging as well.

1. Open **Categories**.
2. Choose **New Category**.
3. Enter **Name** and choose **Color**.
4. Choose **Create Category**.

The **Sort Order** field does not change the displayed category ordering. Use the category names and filters to locate a category instead of relying on that field.

For tag-name changes, use the Tags page. Existing annotations retain their category text rather than automatically following a renamed tag; review their category fields if you need consistent filtering after a rename.

## Troubleshooting

- **An older event appears on today's date:** the annotation records its creation date; include the actual event date in the description.
- **Unexpected category names:** categories are shared project tags and may have been created while organizing prompts.
- **Category count differs from prompt count:** category cards count annotations; the Tags table counts tagged prompts.
- **Create is blocked:** confirm role and plan access with a workspace admin.
- **No marker in a report:** check the report's date range and whether that view supports annotations. The annotation list is the place to confirm the saved note.

## Related guides

- [Tags and assignments](https://docs.equalai.io/tracking/tags)
- [Project settings](https://docs.equalai.io/account/project-settings)
- [Contact support](https://docs.equalai.io/account/support)


---

# Track competitors

Manage project-specific competitors, review suggestions, and choose identifiers that help recognize competing brands in AI answers.

Canonical: https://docs.equalai.io/tracking/competitors
Markdown: https://docs.equalai.io/tracking/competitors.md

Competitors give a project a defined comparison set. Each competitor has a name, domain, chart color, and optional identifiers or aliases. Adding a competitor does not create another project or give anyone access to your workspace.

## Before you edit

Select the intended project. Competitor changes require editing access and usable subscription access. The competitor allowance applies per project; it is different from the workspace's project and tracked-prompt capacity.

Open **Competitors**. The page has **Current Competitors** and **Suggested** tabs. Search by competitor name or domain; the current list also supports matching identifiers.

![EarlySEO Competitors list with Add Competitor, Current and Suggested tabs, and brand matching identifiers marked](https://docs.equalai.io/images/guides/competitors.webp)

*EarlySEO example comparison set. **1.** Add Competitor. **2.** Current or Suggested competitors. **3.** Brand identifiers used for matching. Choose competitors and aliases relevant to your own project.*

## Add a competitor

1. Choose **Add Competitor**.
2. Enter the competitor's **Name** and **Domain**.
3. Choose a **Color** to distinguish it in comparisons.
4. Add relevant **Identifiers and aliases**, if needed.
5. Choose **Save competitor**.

Check that the saved name and domain appear in **Current Competitors**. Review the existing list before adding a domain that is already tracked; a repeated domain is not a reliable way to create a separate comparison entry.

### Choose precise identifiers

Identifiers are additional names that help recognize the competitor in AI answers. Suitable choices include distinctive abbreviations or a strongly associated product name. Avoid a generic category term such as “software” that could describe many unrelated companies.

You can paste comma-separated aliases and remove individual alias chips. The project's **Alternative Names** field and the competitor alias editor are separate settings: editing one does not configure the other.

**Synthetic example:** for a fictional competitor named `Northstar Outdoor`, an alias such as `Northstar Packs` could be useful if it is a real name that brand uses. Do not add a name solely to make more answers count.

## Review suggestions

Open **Suggested** and inspect the name, domain, and explanation before choosing **Add**. A suggestion is a candidate, not an endorsement or a verified direct competitor.

Adding a suggestion uses the same project capacity as manually adding a competitor. If the project is at its limit, review existing competitors or ask a workspace admin to review the plan.

Use the dismiss/reject control to hide an irrelevant suggestion from the current view. It may appear again on a later visit; dismissal should not be treated as a permanent exclusion rule.

## Edit or remove an entry

Use the edit action beside a current competitor to update its identity, aliases, or chart color, then save. Color is a visual distinction, not a measure of performance.

Use the delete action and confirm **Delete Competitor** to remove an entry from the project's current competitor list. Before removing one that your team relies on, agree on the replacement comparison set. Do not assume a configuration change reruns prompts or regenerates historical answers.

## Troubleshooting

- **No current competitors:** add a relevant competitor manually or review suggestions.
- **No search results:** clear the query and confirm the selected project.
- **Wrong brand recognized:** check the domain and remove ambiguous identifiers.
- **Save or add blocked:** confirm editing permissions, subscription access, and the per-project competitor allowance.
- **Results have not changed:** competitor configuration is not itself a prompt run. Review [scheduling and manual runs](https://docs.equalai.io/tracking/schedules).

## Related guides

- [Project identity and alternative names](https://docs.equalai.io/account/project-settings)
- [Organize prompts with tags](https://docs.equalai.io/tracking/tags)
- [Workspace subscription and limits](https://docs.equalai.io/account/billing)


---

# Schedule and run prompts

Configure automatic tracking, run all or selected prompts, and review model, fan-out, timezone, and credit choices before queuing work.

Canonical: https://docs.equalai.io/tracking/schedules
Markdown: https://docs.equalai.io/tracking/schedules.md

The **Schedule and run prompts** dialog separates two actions: saving automatic tracking settings and queuing a manual run. A manual run does not save your schedule changes.

## Before you start

Select the intended project and confirm that it has tracked prompts. You need project editing access, usable subscription access, and access to the models you choose. Running prompts also requires enough AI credits.

Schedules apply to the project's regular tracked prompts, not the onboarding-only prompt set. Selecting a few rows on the prompts page controls a manual selected-prompt run; it does not create a schedule for only those rows.

## Set up automatic runs

1. Open **Schedule and run prompts**.
2. Set **Automatic runs** to **On**.
3. Choose **Weekly** or **Monthly**.
4. For weekly runs, select the weekdays and interval. For monthly runs, choose the day of the month and interval.
5. Choose the models and review any locked add-on models.
6. Review the run time and the timezone displayed beside it.
7. Review the upcoming-run preview.
8. Choose **Save schedule** and wait for **Schedule saved**.

![Schedule and run prompts dialog with automatic runs Off, cadence controls, and selected model credit costs marked](https://docs.equalai.io/images/guides/schedule.webp)

*EarlySEO example setup. **1.** Automatic runs are Off. **2.** Cadence controls. **3.** Selected models and credit costs. Save schedule is disabled in this unchanged example; it does not show an enabled or newly saved schedule. Use the costs displayed in your own dialog.*

Intervals can span 1–12 weeks or months. For a monthly day beyond the end of a shorter month, the run falls on that month's last day.

The time control uses your selected app timezone, while the schedule is stored in UTC. Check the preview after changing your timezone or near a seasonal clock change; do not assume the same displayed local hour will apply all year.

To stop future automatic runs, set **Automatic runs** to **Off** and save. This is not a cancellation control for work already queued.

### Saving does not spend credits

**Save schedule** updates configuration. The credit estimate is a planning aid, not a charge for pressing Save. Actual execution requires available credits when work is queued. A saved schedule therefore does not guarantee that every future run will execute if the balance or model access changes.

## Run all eligible prompts now

Choose the models and fan-out option for this run, then use **Run now**. EqualAI queues eligible prompts from the selected project with those choices.

The manual run uses the choices currently displayed, even if they differ from your saved schedule. To retain those choices for future automatic runs, use **Save schedule** separately.

A success message confirms that work was queued. Follow the active/queued status; it does not promise that all model answers are already complete.

## Run selected prompts

Select prompts on the prompts page before opening the dialog. In **Run selected prompts**, confirm the selected scope, choose models and fan-out, and use the run button showing the selected count.

Only those prompts are submitted. The saved project schedule stays unchanged. If a scope selector offers all prompts instead, check it carefully before submitting: switching scope changes the work and credit estimate.

## Understand model and delivery costs

Each selected model contributes to the cost of a run. Models can have different credit weights; use the per-model cost and total shown in the dialog rather than assuming one credit per model.

**Fan-out** adds phrasing variations alongside the original prompt. The helper shows the number of variations and resulting multiplier. More variations mean more execution work and higher credit use.

**Instant Run** applies the displayed premium credit multiplier to this manual run. Choose **Run instantly** or **Run selected instantly** only after reviewing that estimate. It does not switch the saved schedule to Instant: automatic runs remain on Standard delivery. Faster delivery is not a fixed completion-time guarantee.

## Disabled buttons and run errors

- **No eligible prompts:** add or select regular tracked prompts and reopen the dialog.
- **No selected models:** choose at least one available model.
- **Locked model:** ask a workspace admin about the required plan or active add-on.
- **Cooldown/countdown:** wait until the displayed delay ends before another run.
- **Not enough credits:** reduce models or fan-out, narrow the selected prompts, or review workspace usage. Instant Run may be blocked even when Standard is affordable.
- **Save schedule disabled:** there may be no unsaved change; a model selection is also required.
- **Another teammate just ran prompts:** the status or balance can change between opening and submitting. Review refreshed values before retrying.
- **Queue or connection error:** do not assume the run started. Check status and contact support if it persists.

## Related guides

- [Workspace credits and capacity](https://docs.equalai.io/account/billing)
- [App timezone](https://docs.equalai.io/account/account-settings)
- [Initial tracking setup](https://docs.equalai.io/getting-started/onboarding)
- [Support](https://docs.equalai.io/account/support)


---

# Organize prompts with tags

Create project tags, import a simple CSV, assign tags to prompts, and understand what the AI suggestion action actually does.

Canonical: https://docs.equalai.io/tracking/tags
Markdown: https://docs.equalai.io/tracking/tags.md

Tags are reusable labels for organizing a project's prompts. A prompt can have several tags. Tags are also used as annotation categories, so treat tag names as shared project vocabulary rather than private labels.

## Create a tag manually

With the intended project selected, open **Tags**:

1. Choose **Create Tag → Create manually**.
2. Enter a required **Name**.
3. Optionally add a **Description** explaining when teammates should use it.
4. Choose **Save Tag**.
5. Return with **Back to tags**.

Creating a tag does not assign it to any prompt. The list shows **Name**, **Description**, and **Prompts**, so a newly created tag can legitimately have a count of zero. Search matches tag names and descriptions.

![EarlySEO Tags list showing tag creation and CSV import, search, and prompt counts](https://docs.equalai.io/images/guides/tags.webp)

*EarlySEO example tags. **1.** Create a tag or import CSV. **2.** Search existing tags. **3.** Prompt count for each tag. A listed tag and its prompt assignments are separate; your names and counts will differ.*

**Synthetic example:** a fictional retailer could use `Buying guide` with the description `Questions about choosing equipment`. Keep labels concise and distinguish them from the full text of a tracked prompt.

## Assign tags to a prompt

Open the prompt's **Add tags** control or existing tag-count control.

1. In **Manage tags**, search for the tags you need.
2. Select or deselect the entries.
3. Choose **Save** to apply the selection.

**Clear tags** clears the pending selection; choose **Save** to apply that change. **Cancel** closes the picker without applying it. If no tags are available, create them in Tags first.

The picker changes the selected prompt, not every prompt in the project. A tag definition and a prompt's tag assignment are separate things.

## Import tags from CSV

Choose **Import tags from CSV** and select a simple CSV file. The first nonempty row must be a header. Use the first column for the name and the second for the description.

This synthetic example contains no customer data:

```csv
name,description
Buying guide,Questions about choosing equipment
Product comparison,Questions comparing alternatives
```

Import up to 500 new tags in one batch. Keep both fields free of embedded commas; quoted CSV fields with commas are not supported by this importer. Do not use a multiline description. Additional columns are not imported.

Blank names and names already present in the project are skipped, with existing-name matching ignoring case. Review the table after import; a successful upload does not mean every source row created a new tag. The importer creates tag definitions, not prompt assignments.

## Use Create with AI

**Create Tag → Create with AI** opens **Create Tags with AI**. It requires at least one prompt; the generate action is disabled when the prompt count is zero.

**Generate Tags with AI** creates a content task titled **Generate tags with AI** to review tag suggestions. It does not automatically create tags or assign them to prompts. Treat the task as a review workflow: decide which labels are useful, create those labels in Tags, and assign them to prompts separately.

The return to the tag list after submitting the task is not confirmation that new tags were added.

## Edit and delete

Use a row's **Edit** action to update the tag, then **Save Tag**. Use **Delete** and confirm **Delete Tag** to remove the tag definition. Removing a label is not the same as deleting a tracked prompt.

Annotation categories share this tag list, but existing annotations retain their category text. After renaming a category tag, review older annotations if you want their category labels to stay consistent.

## Troubleshooting

- **No tags yet:** create one manually or import a correctly formatted CSV.
- **No matching tags:** clear search before assuming the project is empty.
- **AI action unavailable:** add a prompt, then reopen the page.
- **Import loses text:** remove embedded commas and verify the two-column format.
- **First tag missing:** check that the file starts with a header row.
- **Save denied:** confirm that you have project editing access and the required subscription access.

## Related guides

- [Annotations and shared categories](https://docs.equalai.io/tracking/annotations)
- [Create your first tracking set](https://docs.equalai.io/getting-started/onboarding)
- [Project roles](https://docs.equalai.io/account/team-members)