> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vantr.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Use the docs with AI

> Connect Vantr developer documentation to AI tools with Markdown exports, llms.txt, skill.md, and the hosted docs MCP server.

# Use the docs with AI

Vantr developer docs are built for people and AI tools. Use the same published docs site for page Markdown, the OpenAPI contract, the hosted search MCP server, and the agent skill.

<Info>
  Set `DOCS_URL` to the published Vantr Developer Docs root before using the examples on this page.
</Info>

<Columns cols={2}>
  <Card title="OpenAPI contract" icon="braces" href="/openapi/v2.json" horizontal>
    Generate clients and verify endpoint behavior from the committed v2 spec.
  </Card>

  <Card title="Docs MCP server" icon="plug-zap" href="/mcp" horizontal>
    Connect an AI tool to search and retrieve current docs content as Markdown.
  </Card>

  <Card title="Documentation index" icon="list-tree" href="/llms.txt" horizontal>
    Give agents a compact directory of available pages and descriptions.
  </Card>

  <Card title="Agent skill" icon="badge-check" href="/skill.md" horizontal>
    Give agents product-specific constraints for building Vantr v2 API integrations.
  </Card>
</Columns>

## One-click context

Use this prompt when configuring an AI coding tool for Vantr API work.

<Prompt description="Give an AI coding tool current Vantr v2 API context." actions={["copy", "cursor"]} icon="sparkles">
  Set DOCS\_URL to the Vantr Developer Docs root.

  Use these documentation sources when building an Vantr v2 integration:

  * OpenAPI contract: \$DOCS\_URL/openapi/v2.json
  * Search MCP server: \$DOCS\_URL/mcp
  * Documentation index: \$DOCS\_URL/llms.txt
  * Full Markdown corpus: \$DOCS\_URL/llms-full.txt
  * Agent skill: \$DOCS\_URL/skill.md

  Install the skill with:

  npx skills add \$DOCS\_URL

  Use only endpoints, scopes, request fields, and response fields documented in the v2 OpenAPI contract or linked guide pages. Keep client secrets server-side.
</Prompt>

## What each endpoint is for

<Tabs>
  <Tab title="Markdown" icon="file-text">
    Every page is available as Markdown by appending `.md` to its URL. Use this when an AI tool needs one page as clean source context.

    ```text Example theme={"theme":{"light":"github-light","dark":"github-dark"}}
    $DOCS_URL/authentication.md
    $DOCS_URL/guides/errors.md
    ```
  </Tab>

  <Tab title="llms.txt" icon="list-tree">
    Use `/llms.txt` as the compact page index and `/llms-full.txt` when an agent needs the complete public docs corpus.

    ```text Sources theme={"theme":{"light":"github-light","dark":"github-dark"}}
    $DOCS_URL/llms.txt
    $DOCS_URL/llms-full.txt
    ```
  </Tab>

  <Tab title="MCP" icon="plug-zap">
    Use `/mcp` for AI tools that support Model Context Protocol. The server exposes docs search and full-page retrieval tools, so answers stay grounded in the current published docs.

    ```text Server URL theme={"theme":{"light":"github-light","dark":"github-dark"}}
    $DOCS_URL/mcp
    ```
  </Tab>

  <Tab title="skill.md" icon="badge-check">
    Use `/skill.md` when an agent needs a concise capability summary and operating rules for Vantr v2 API work.

    ```text Skill URL theme={"theme":{"light":"github-light","dark":"github-dark"}}
    $DOCS_URL/skill.md
    ```
  </Tab>
</Tabs>

## Recommended agent workflow

<Steps titleSize="h3">
  <Step title="Start from the spec" icon="braces">
    Read `/openapi/v2.json` before writing client code, request payloads, or endpoint mappings.
  </Step>

  <Step title="Read the guide page" icon="book-open-check">
    Pair the API reference with the guide for authentication, scopes, conventions, errors, and the resource workflow you are building.
  </Step>

  <Step title="Use the contextual menu" icon="sparkles">
    Copy a page as Markdown, open it as `.md`, send it to ChatGPT or Claude, download the spec, or copy the docs MCP server URL from the page header.
  </Step>

  <Step title="Keep secrets server-side" icon="lock-keyhole">
    Treat client secrets and bearer tokens as server-only credentials. Never place them in browser code, mobile apps, logs, or shared prompts.
  </Step>
</Steps>

Agent instructions for Vantr v2 API work:

* Prefer `/openapi/v2.json` as the endpoint and schema contract.
* Use `/authentication`, `/guides/scopes`, `/guides/api-conventions`, `/guides/errors`, and resource-specific guide pages before generating code.
* Use only documented v2 endpoints, scopes, request fields, response fields, and enum values.
* Keep client secrets and bearer tokens server-side.
* For generated clients, include structured non-2xx response handling and preserve pagination fields.

## Related pages

<Columns cols={3}>
  <Card title="Generate clients" icon="package-check" href="/guides/client-generation">
    Generate typed clients from the published OpenAPI spec.
  </Card>

  <Card title="Authentication" icon="key-round" href="/authentication">
    Choose Basic auth, client headers, or OAuth bearer tokens.
  </Card>

  <Card title="API conventions" icon="waypoints" href="/guides/api-conventions">
    Understand pagination, envelopes, casing, and archive behavior.
  </Card>
</Columns>