For the complete BrainAPI documentation index, see llms.txt. A markdown version of any docs page is available by appending .md to its URL. Docs MCP: /docs/mcp.
BrainAPI Docs MCP
Connect coding agents to BrainAPI documentation over streamable HTTP
This is the documentation MCP — search and fetch public docs as markdown. It is separate from the product MCP that exposes BrainAPI graph tools on your API host (typically :8001).
Endpoint: https://brain-api.dev/mcp
Discovery: https://brain-api.dev/.well-known/mcp.json
Machine index: /docs/llms.txt · append .md to any docs URL for markdown
Tools
| Tool | Purpose |
|---|---|
search_docs | Full-text search (tag: v1 or v2) |
list_sections | List pages by version (default v2) |
get_page | Fetch a page as markdown |
get_code_example | Extract the first (or language-matched) fenced code block |
Cursor
Add to .cursor/mcp.json (or Cursor Settings → MCP):
{
"mcpServers": {
"brainapi-docs": {
"url": "https://brainapi.lumen-labs.ai/docs/mcp"
}
}
}Claude Desktop / Claude Code
If your client supports streamable HTTP MCP natively, point it at the same URL. Otherwise use a stdio→HTTP adapter:
{
"mcpServers": {
"brainapi-docs": {
"command": "npx",
"args": ["-y", "@pyroprompts/mcp-stdio-to-streamable-http-adapter"],
"env": {
"URI": "https://brainapi.lumen-labs.ai/docs/mcp",
"MCP_NAME": "brainapi-docs"
}
}
}
}No BrainPAT is required — this server only serves public documentation.
Agent workflow
- Prefer V2 (
search_docswithtag: "v2", or start from /docs/llms/v2.txt). get_pagewith a path like/v2/ingestion/tasks.- For API calls: expect ingest 202 +
task_id; pollGET /tasks/{id}; 404 means unknown task, not pending. - Optional skill summary: /docs/skill.md
Related
- Product MCP — tools against a live BrainAPI instance
- llms.txt — short TOC for agents
- Ingestion tasks — async poll contract
Last updated on
