KnitKnot

# Check whether AI agents can read your site

Read the Content Access checks, open a failing domain or page finding, and tell a site problem from a scanner failure.

## At a glance

  1. 1. **Find a failing check**: Open Content → Access and filter the domain, status, or check.
  2. 2. **Read the evidence**: Open a row to inspect the finding, affected pages, and the last check time.
  3. 3. **Recheck after a fix**: Run the checks again and confirm the new status before interpreting traffic or citations.

Open [Content → Access](https://app.knitknot.ai/content/access) to see whether KnitKnot could reach and inspect your site. The checklist covers domain checks such as `robots.txt`, AI crawler access, `llms.txt`, sitemap health, organization schema, and snippet directives. It also includes page findings for unreadable, retrying, or cited pages that no longer load.

A passing check says what the scanner observed when it ran. It does not promise that an AI engine will cite the page. For actual agent requests, use [Agent activity](/docs/read-ai-traffic/); for what an engine said in an answer, open the [benchmark](/docs/read-your-report/).

## Find the problem

  1. 1. Open **Content → Access**. Checks with failures appear ahead of informational and passing rows until you choose a different sort.
  2. 2. Search for a domain or finding, or filter by **Domain**, **Status**, or **Check**. Read the **Finding** text; it contains the specific condition and suggested next move.
  3. 3. Open a row. The detail page shows the status, when the domain was checked, and the relevant evidence. The **AI crawler access** detail can also show a verdict for each supported engine when that evidence is available.
  4. 4. If the finding names affected pages, open those page rows before changing a site-wide setting. A failed page fetch and a domain-wide crawler rule call for different fixes.

For example, if one cited URL now returns an error, check that URL and its redirect before editing `robots.txt`. If the AI crawler access check reports a domain rule, inspect that rule and the per-engine verdicts before asking your site owner to change it.

## Tell a failed check from a scan that could not run

A **failed** check reports a condition the scanner observed on the site. **Some checks couldn't run** means a firewall, bot protection, or another fetch problem prevented the scanner from completing the check. It is not evidence that the page itself failed the test. The warning appears above the table when a fetch error exists.

Select **Re-run checks** after changing the site or when a fetch error may have cleared. While a domain is still being checked, Access shows its in-progress state. Wait for the result before treating an earlier status as current. A detail page also offers the recheck action when its check has a fetch error.

If Access says **No crawl domains yet**, [add a website source](/docs/manage-authority-sources/) under **Facts → Authority**. The checklist starts from domains KnitKnot has crawled; an empty table does not mean every crawler can reach your site.

## What to check next

After a site fix, read the latest check and then inspect [Agent activity](/docs/read-ai-traffic/) for verified requests. Requests show access, while captured [answers and citations](/docs/investigate-a-claim/) show what an AI system returned. Keep those observations separate when reporting whether the fix helped.

Docs navigation
Docs › Guides

Check whether AI agents can read your site

Read the Content Access checks, open a failing domain or page finding, and tell a site problem from a scanner failure.

Updated

At a glance

  1. Find a failing check

    Open Content → Access and filter the domain, status, or check.

  2. Read the evidence

    Open a row to inspect the finding, affected pages, and the last check time.

  3. Recheck after a fix

    Run the checks again and confirm the new status before interpreting traffic or citations.

Open Content → Access to see whether KnitKnot could reach and inspect your site. The checklist covers domain checks such as robots.txt, AI crawler access, llms.txt, sitemap health, organization schema, and snippet directives. It also includes page findings for unreadable, retrying, or cited pages that no longer load.

A passing check says what the scanner observed when it ran. It does not promise that an AI engine will cite the page. For actual agent requests, use Agent activity; for what an engine said in an answer, open the benchmark.

Find the problem

  1. Open Content → Access. Checks with failures appear ahead of informational and passing rows until you choose a different sort.
  2. Search for a domain or finding, or filter by Domain, Status, or Check. Read the Finding text; it contains the specific condition and suggested next move.
  3. Open a row. The detail page shows the status, when the domain was checked, and the relevant evidence. The AI crawler access detail can also show a verdict for each supported engine when that evidence is available.
  4. If the finding names affected pages, open those page rows before changing a site-wide setting. A failed page fetch and a domain-wide crawler rule call for different fixes.

For example, if one cited URL now returns an error, check that URL and its redirect before editing robots.txt. If the AI crawler access check reports a domain rule, inspect that rule and the per-engine verdicts before asking your site owner to change it.

Tell a failed check from a scan that could not run

A failed check reports a condition the scanner observed on the site. Some checks couldn’t run means a firewall, bot protection, or another fetch problem prevented the scanner from completing the check. It is not evidence that the page itself failed the test. The warning appears above the table when a fetch error exists.

Select Re-run checks after changing the site or when a fetch error may have cleared. While a domain is still being checked, Access shows its in-progress state. Wait for the result before treating an earlier status as current. A detail page also offers the recheck action when its check has a fetch error.

If Access says No crawl domains yet, add a website source under Facts → Authority. The checklist starts from domains KnitKnot has crawled; an empty table does not mean every crawler can reach your site.

What to check next

After a site fix, read the latest check and then inspect Agent activity for verified requests. Requests show access, while captured answers and citations show what an AI system returned. Keep those observations separate when reporting whether the fix helped.