ClaudeChatGPTPerplexityGeminiMicrosoft CopilotRaycastMeta AIGrokZ.aiQwenKimi
DeepSeekMistralCursorVS CodeWindsurfJetBrainsClineLovableVercel AI SDKLangChain

Use DocBreach with your AI.

Connect your account once and let the AI you already use work with it, without building another integration. Real-time documentation discovery for AI agents. Discover, read, and extract any API documentation without browser rendering or API keys.

Included with plan

Ask AI about this Connector

Developed, maintained, and hosted by Vinkius.

MCP VERIFIED · PRODUCTION READY · VINKIUS GUARANTEED

Waiting for input…

Works with modern AI clients that support MCP, including ChatGPT, Claude, Cursor, and more.

ChatGPTClaudeCursorPerplexityGeminiMicrosoft CopilotRaycastMeta AI

Complete set · 5 capabilities

The complete DocBreach capability set.

These are the exact actions your AI can choose when you ask it to work with DocBreach.

Capability set01 / 02

01-03

3 capabilities in this set.

Part of 5 available through DocBreach.

  1. 01

    Docs map

    Provide a domain (e.g., "stripe.com") and receive a complete table of contents with every documentation page organized by section. Then use docs.read on specific pages from the map. Map the complete documentation structure of any domain

  2. 02

    Docs read

    Handles HTML, JSON, YAML, OpenAPI specs, Postman Collections, PDFs (<5MB), and llms.txt. The response includes a "Related Documentation Links" section extracted from page navigation. ALWAYS check these links for authentication and getting-started pages before generating code. Read any documentation URL and return clean, LLM-ready Markdown

  3. 03

    Docs search

    Always provide the "site" parameter when searching within a known domain. Example: docs.search({ query: "authentication headers", site: "docs.stripe.com" }) Search for specific topics within a documentation site

Capability set02 / 02

04-05

2 capabilities in this set.

Part of 5 available through DocBreach.

  1. 04

    Docs discover

    Use specific, descriptive queries. Example: "stripe API webhooks" instead of just "stripe". Combine your search intents into a single query. Do NOT call this capability in rapid loops. refine your query instead. Find documentation sources for any service, library, or API

  2. 05

    Docs extract

    Read has identified an OpenAPI/Swagger spec URL or Postman Collection. Provide a tag to filter endpoints by API group. If no tag is provided, returns a summary of all tags with endpoint counts. Extract structured endpoint information from an OpenAPI, Swagger, or Postman spec

Observed, not estimated

1295ms average. Fast in production.

DocBreach is checked daily against the live service.

Daily averagePeak 1434ms
Aug 23Today
Fastest day
1022ms
Slowest day
1434ms
14-day trend
Stable-3%

Connect your client

One URL. Every client.

Activate the Connector, copy your link, and paste it into the client you already use. 5 capabilities arrive ready to run.

Preview access · not provider authentication

The vk_preview_* token belongs to Vinkius preview infrastructure. It lets Claude discover and display the capabilities of DocBreach, so you can see the experience inside your AI.

It does not authenticate your account with DocBreach. Actions requiring credentials or live account data may not run until you activate the Connector and authorize the service.

DocBreach Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_irgtI6t1CMNMo7SoPiRQBNgk7Rz1GQFQBkyR1OUW/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — DocBreach capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "docbreach-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_irgtI6t1CMNMo7SoPiRQBNgk7Rz1GQFQBkyR1OUW/mcp"
    }
  }
}
  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Step-by-step instructions for each client are in the guide. How to connect

FAQ

Questions DocBreach owners ask.

  • 01

    How does DocBreach bypass modern Single Page Application (SPA) barriers?

    DocBreach features a built-in hydration engine. Rather than relying on a heavy headless browser, it intercepts and evaluates underlying framework payloads (like __NEXT_DATA__ for Next.js or Docusaurus state) to extract the raw documentation text directly. This makes it blisteringly fast and resource-efficient.

  • 02

    What documentation formats and frameworks are officially supported?

    It natively parses OpenAPI/Swagger specifications, Postman Collections, and standard llms.txt files. For web-based docs, it surgically cleans noise from 12+ major frameworks including Docusaurus, Nextra, VitePress, Mintlify, GitBook, and ReadMe, returning LLM-optimized Markdown.

  • 03

    Do I need any third-party API keys or paid scraping subscriptions?

    Absolutely not. DocBreach is designed to operate completely independently. It accesses public documentation directly via optimized HTTP clients using smart resolution strategies, eliminating the need for proxy services, SaaS subscriptions, or scraping API keys.

  • 04

    How does this MCP server improve my AI Agent's performance?

    AI Agents typically hallucinate when APIs are updated or when facing undocumented endpoints. By equipping your agent with DocBreach, it can autonomously search, map, and read the definitive, up-to-date source of truth before writing a single line of code, ensuring deterministic integrations.

  • 05

    Can it extract specific endpoints from massive OpenAPI specifications?

    Yes! The docs.extract capability is built precisely for this. Instead of loading a 5MB JSON spec into your agent's context window (which wastes tokens and ruins attention), DocBreach parses the spec server-side and allows your agent to query exactly the endpoints and methods it needs.