# Check your llms.txt file, line by line

Canonical: https://llmcitely.com/tools/llms-txt-validator
Publisher: LLM Citely
Reviewed: 2026-10-31

Paste an llms.txt file to review headings, file links and URL syntax. See concrete fixes and download the supplied file and review record locally.

![One required H1. Useful optional guidance. The v2 llms.txt proposal describes a small Markdown directory.](https://llmcitely.com/images/editorial/llms-required-and-optional.svg)

The v2 llms.txt proposal describes a small Markdown directory. A format review does not grant crawler access, verify content or guarantee an AI citation. Original explanatory diagram; illustrative counts and states are not customer measurements.

Interactive tool: [open the browser version](https://llmcitely.com/tools/llms-txt-validator#resource-interactive). This text edition describes the workflow; it does not execute the tool or contain a visitor's inputs or results.

## Review the supplied file

Paste the Markdown content you plan to serve as llms.txt. Add its eventual file URL if you want relative links resolved. Choose Check llms.txt to inspect the file; the browser does not request that URL.

The examples demonstrate both a concise valid file and a document with issues. They contain example.com links and are not live-access tests. Changing an input clears the previous review so an export cannot silently describe an earlier file.

The tool accepts up to 512 KiB and 4,000 lines. It checks the common Markdown list-link form and flags complex items for a manual review rather than treating its parser as every possible Markdown implementation.

## Keep the required H1 separate from suggestions

In the current [llms.txt proposal](https://llmstxt.org/), the site-name H1 is the only required section. A blockquote summary, explanatory text and H2 file-list sections can make a directory more useful, but they are not all mandatory.

| Finding | Interpretation | Example response |
|---|---|---|
| Missing or repeated H1 | Changes needed | Keep one project-name H1 at the start |
| Relative file link without a base | Review needed | Supply the file URL or use an absolute URL |
| Missing short summary | Optional suggestion | Add a concise description when useful |
| Parsed link | Syntax observed | Still verify the actual served resource separately |

A short `# Project name` file can pass the required format checks and still benefit from a useful directory. The checker does not turn optional recommendations into a percentage score.

## URL syntax is not an HTTP check

For each parsed file link, inspect its line, label, section and resolved destination. The checker rejects non-HTTP schemes and URLs containing credentials. It reports duplicate destinations and file links outside the expected section structure for review.

Every network-access entry remains **Not checked**. A valid URL can redirect, require a login, return an error or contain inaccurate content. Test those facts separately before publishing.

Use the [HTML checker](https://llmcitely.com/tools/ai-readiness-checker) for supplied-page declarations and the [crawler-rule checker](https://llmcitely.com/tools/ai-crawler-checker) for a specific path and robots file. Neither is replaced by a successful llms.txt format check.

## Publish a small directory you can maintain

Save the review JSON with your implementation notes. The llms.txt download contains the supplied text with normalized line endings; it does not silently rewrite or fix your content. Apply the suggested edits, then check the new file again.

After publishing, inspect the actual served file, its referenced pages and any Markdown alternatives. Keep the directory focused and remove stale links when a page moves. The current proposal recommends discoverable text alternatives and links to the relevant llms.txt directory.

Google's [AI-feature guidance](https://developers.google.com/search/docs/appearance/ai-features) does not require a new AI text file or special markup to appear in its AI experiences. Maintain useful public HTML and truthful content; the file is an additional navigation aid, not a citation or ranking guarantee.

## Questions, answered

### Does the checker fetch my links?

No. It checks supplied text and resolves local URL syntax. HTTP status, actual access and destination content remain untested.

### Do I need every suggested section to pass?

No. The proposal requires the site-name H1. Other findings include optional suggestions and manual-review notes.

### Does the download contain an automatically repaired file?

No. It contains the supplied text with normalized line endings. Apply the suggestions yourself and run a new review.

## Sources and further reading

- [The llms.txt proposal, v2](https://llmstxt.org/)
- [Google: AI features and website requirements](https://developers.google.com/search/docs/appearance/ai-features)

Examples marked illustrative are not customer measurements.

## Continue with the next step

- [Review crawler rules separately](https://llmcitely.com/tools/ai-crawler-checker)
- [Review supplied page HTML](https://llmcitely.com/tools/ai-readiness-checker)
- [Understand access and training controls](https://llmcitely.com/guides/robots-txt-for-ai-crawlers)
- [Keep the checks in a scoped audit](https://llmcitely.com/guides/run-a-geo-audit)

## Your next step

[Review crawler rules separately](https://llmcitely.com/tools/ai-crawler-checker)