# 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).