> For the complete documentation index, see [llms.txt](https://learning.contextqa.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://learning.contextqa.com/reference/using-contextqa-docs-with-ai.md).

# Use the Docs with Search and AI

## Quick answer

The public ContextQA documentation is available to both search crawlers and AI retrieval tools. Search engines can index the canonical HTML pages through the site's sitemap. AI agents can discover the documentation through `llms.txt`, retrieve a compact Markdown version of any page, or ask GitBook a focused question and receive an answer grounded in the published documentation.

Use the HTML URL when sharing or citing a page for people. Use `llms.txt` and the `.md` URLs when supplying ContextQA documentation to an AI agent or retrieval pipeline.

## Verified public retrieval signals

The published documentation currently exposes all of the following:

* `robots.txt` allows the public site to be crawled and publishes `search=yes` and `ai-input=yes` content signals.
* `sitemap.xml` points to the page sitemap, where each canonical page has a modification timestamp.
* `llms.txt` lists page titles, descriptions, and machine-readable Markdown links.
* Each HTML page is indexable and includes a unique title, description, and canonical URL.
* Each published page has a `.md` representation with the main content and links but without the visual site chrome.
* GitBook's `ask` and optional `goal` parameters let an agent request an answer grounded in the published documentation.

These signals make the content discoverable and easy to retrieve. They do not force Google, ChatGPT, Claude, or another product to index, rank, quote, or cite a page; each provider controls those decisions.

## Public discovery endpoints

| Purpose                      | Endpoint                                                  | What it provides                                         |
| ---------------------------- | --------------------------------------------------------- | -------------------------------------------------------- |
| Search crawler policy        | [robots.txt](https://learning.contextqa.com/robots.txt)   | Crawl permissions and the sitemap location               |
| Search index                 | [sitemap.xml](https://learning.contextqa.com/sitemap.xml) | Canonical documentation URLs and modification dates      |
| AI documentation index       | [llms.txt](https://learning.contextqa.com/llms.txt)       | Page titles, descriptions, and links to Markdown sources |
| Human-readable documentation | `https://learning.contextqa.com/<page>`                   | Canonical HTML page for readers and search results       |
| Machine-readable page        | `https://learning.contextqa.com/<page>.md`                | Compact Markdown containing the page text and links      |

For example:

* Human-readable page: [AI Insights](https://learning.contextqa.com/ai-features/ai-insights)
* Markdown source: [AI Insights Markdown](https://learning.contextqa.com/ai-features/ai-insights.md)

The Markdown endpoints are retrieval sources, not separate pages that need to rank in search. When an agent cites a source for a reader, prefer the corresponding canonical HTML URL without `.md`.

## Give the documentation to an AI agent

Use this retrieval order so the agent receives current, relevant context without loading the entire site:

1. Start with `https://learning.contextqa.com/llms.txt` to discover the available pages.
2. Select the pages whose titles and descriptions match the question.
3. Retrieve those pages by appending `.md` to their canonical URLs.
4. Answer only from the retrieved text and preserve links to the canonical HTML sources.
5. If the answer is incomplete, query the documentation dynamically before making an assumption.

Example instruction for an AI tool:

```
Use https://learning.contextqa.com/llms.txt as the documentation index.
Retrieve the relevant .md pages, answer from those sources, and cite the
matching canonical HTML URLs without the .md suffix. If the documentation
does not support a claim, say that the detail is not documented.
```

## Use the documentation with ChatGPT, Claude, and other assistants

When an assistant can open public URLs, give it the documentation index and a precise question. This works with browser-enabled or retrieval-enabled assistants from different providers because it relies on ordinary public HTTP pages rather than a provider-specific format.

Reusable prompt:

```
Use https://learning.contextqa.com/llms.txt to find the most relevant
ContextQA pages. Retrieve their .md versions, answer only from those pages,
and cite the matching canonical HTML URLs without .md. Separate confirmed
product behavior from workspace-, permission-, or deployment-specific details.
If the sources do not answer the question, say what is not documented.
```

For a single known topic, link the canonical page directly instead of asking the assistant to search the entire site. For example, use the [Test Plans page](https://learning.contextqa.com/web-testing/test-plans) for questions about plan creation, capacity, live execution, queues, and stopping a run.

## Ask the documentation a focused question

GitBook can answer a question using the published ContextQA documentation. Add an `ask` query parameter to a Markdown page URL:

```
https://learning.contextqa.com/readme.md?ask=How%20does%20ContextQA%20generate%20tests%20from%20Jira%3F
```

You can also add an optional `goal` parameter to explain the broader task:

```
https://learning.contextqa.com/readme.md?ask=How%20do%20I%20run%20a%20test%20plan%3F&goal=Create%20a%20CI%2FCD%20quality%20gate
```

Questions should be specific and self-contained. Treat the returned answer as a retrieval aid: retain its source links and verify security-sensitive, permission-sensitive, or deployment-specific instructions against the referenced page.

## Help search engines understand a page

When linking to ContextQA documentation from the main website, a release note, GitHub, or an integration listing:

* Link to the canonical HTML page that answers the question, not only the documentation home page.
* Use descriptive anchor text such as “configure ContextQA Jira integration,” not “click here.”
* Link related concepts bidirectionally—for example, connect API chaining to variables, environments, and response validation.
* Keep page titles and opening answers aligned with the language customers use when asking questions.
* Update the existing page when a workflow changes instead of publishing a second competing answer.

## Repository-wide citation-ready page contract

Every published page follows the same retrieval contract so a reader or AI agent can extract the main answer without reconstructing it from screenshots:

1. Use one specific page title and one matching level-one heading.
2. Write a unique description that names the task, product object, and outcome.
3. Put a **Quick answer** near the top with the shortest complete answer to the page's main question.
4. Use exact visible labels and stable terms, such as **Runs → Live execution**, then define what each value means.
5. Separate prerequisites, steps, verification, limitations, and troubleshooting under descriptive headings.
6. State when behavior depends on permissions, subscription, deployment, retention, or artifact availability.
7. Describe screenshots with the page, state, and control the reader should notice; never make the screenshot the only source of an instruction.
8. Link to the most specific related page with descriptive anchor text.
9. Cite the canonical HTML URL when presenting an answer to a person, even if the agent retrieved the `.md` representation.

The repository documentation audit applies this contract to every page in the published navigation. It enforces unique titles, unique discovery descriptions, one level-one heading, valid heading hierarchy, a concise near-top **Quick answer**, at least one related documentation link, valid local links and assets, descriptive image text, and canonical demo links. A new page cannot pass the audit by copying another page's title, description, or answer.

## Source and citation guidance

For the most reliable answer, prefer sources in this order:

1. A feature-specific or task-specific ContextQA documentation page.
2. A ContextQA tutorial that demonstrates the workflow end to end.
3. A section overview or role guide.
4. The documentation changelog for documentation-history questions.

The documentation describes user-facing behavior. Availability can differ by workspace permissions, subscription, or deployment. Do not infer unpublished APIs, internal implementation details, credentials, or business rules from screenshots or portal network calls.

## Frequently asked questions

<details>

<summary>Does ContextQA publish an llms.txt file?</summary>

Yes. [ContextQA llms.txt](https://learning.contextqa.com/llms.txt) lists the published documentation pages with descriptions and Markdown links so an AI retrieval tool can select relevant sources.

</details>

<details>

<summary>Can I retrieve a ContextQA documentation page as Markdown?</summary>

Yes. Append `.md` to the canonical page URL. For example, `/api-testing/api-chaining` becomes `/api-testing/api-chaining.md`.

</details>

<details>

<summary>Which URL should an AI answer cite?</summary>

Retrieve the Markdown page when that format is easier to process, but cite the canonical HTML URL without `.md` for readers.

</details>

<details>

<summary>Does llms.txt guarantee that every AI product will index or cite ContextQA?</summary>

No. It makes discovery and retrieval easier for tools that use it, but each search engine and AI product controls its own crawling, indexing, ranking, and citation behavior. Clear answers, accurate content, relevant external links, and consistent updates still matter.

</details>

<details>

<summary>Can an AI agent operate ContextQA as well as read the documentation?</summary>

Yes, after authentication and setup. Reading the documentation is a public retrieval workflow; operating a workspace uses the [ContextQA MCP Server](/mcp-server/overview.md) and its authenticated tools.

</details>

## Related pages

* [ContextQA documentation overview](/readme.md)
* [MCP Server overview](/mcp-server/overview.md)
* [Agent integration guide](/mcp-server/agent-integration-guide.md)
* [Glossary](/reference/glossary.md)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://learning.contextqa.com/reference/using-contextqa-docs-with-ai.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.