Skip to main content
AI applications can access the TechWolf Developer Portal documentation programmatically, enabling AI agents to answer questions about the Skill Engine API, generate integration code, and even execute API calls on behalf of users. There are three ways to connect AI applications to our documentation, each suited for different use cases.

​
MCP Server

The Model Context Protocol (MCP) server allows AI applications to search and query the TechWolf Developer Portal in real-time. This is the recommended approach for interactive AI workflows. URL: https://developers.techwolf.ai/mcp

​
Setup

Add the following to your MCP client configuration (Claude Code, Cursor, Windsurf, etc.): 
{
  "mcpServers": {
    "techwolf": {
      "type": "url",
      "url": "https://developers.techwolf.ai/mcp"
    }
  }
}

​
Available tools

ToolDescription
SearchTechWolfDeveloperPortalSearch across the knowledge base for documentation, code examples, API references, and guides.

​
When to use

  • Building AI-powered integrations that need to look up API documentation on-the-fly.
  • AI coding assistants (Cursor, Windsurf, Claude Code) that need context about the Skill Engine API while writing integration code.

​
llms-full.txt

A single file containing the full content of the TechWolf Developer Portal. Load this into an AI application’s context window for comprehensive, one-shot access to all documentation. URL: https://developers.techwolf.ai/llms-full.txt

​
When to use

  • You need the AI to have complete knowledge of the documentation upfront.
  • Your AI tool does not support MCP.
This file is very large. Ensure your AI application’s context window can accommodate it or use subagents.

​
llms.txt

An AI-friendly sitemap that lists all documentation pages with their URLs. This helps AI systems discover and index the TechWolf Developer Portal content. URL: https://developers.techwolf.ai/llms.txt

​
When to use

  • AI crawlers or indexing systems that need to discover available documentation.
  • You want the AI to selectively fetch only the pages relevant to a specific question, rather than loading everything at once.
  • Lighter alternative when full context loading is not practical.