Overview
K-AI exposes two Model Context Protocol (MCP) servers — one for Retrieval (5 tools, the canonical replacement of the /search endpoint that was removed from the K-AI Instance API in April 2026 — see Removed endpoints) and one for Audit (20 curated tools). MCP clients (Claude Desktop, Cursor, Mistral Le Chat, any MCP-compatible agent) discover OAuth at each MCP URL and authenticate the end-user once. A single token grants access to every K-AI instance the user is entitled to, on both servers. The host LLM does the orchestration; K-AI exposes primitives, not a chat strategy.
Two MCP servers
https://api-retrieval.kai-studio.ai/mcp— 5 tools. The Activate surface over Retrieval. See Retrieval tools.https://api-audit.kai-studio.ai/mcp— 20 tools. The Activate surface over K-AI Audit, organised conflict-first. See Audit tools. Both servers follow the schema stability policy.
Both servers use Streamable HTTP transport and are stateless (no server-side sessions).
Authentication
Both surfaces share a single OAuth 2.1 Authorization Server: https://auth-api.kai-studio.ai. Clients discover it via /.well-known/oauth-authorization-server on either MCP host. Tokens are signed JWTs and are accepted on both surfaces — register your client once, use it everywhere. End-user identity is propagated into every tool call, carrying group-based RBAC, per-instance visibility and instructions. See OAuth 2.1 for the full flow.
Quick start
The fastest path is Cursor: two one-click install buttons (one per server). Claude Desktop and Mistral Le Chat require UI steps via their respective Custom Connector panels — see Connect a client.
What's in this section
Protocol primer; why K-AI exposes MCP.
Install in Claude Desktop, Cursor, Mistral Le Chat.
Sequence diagram, endpoints, troubleshooting.
The 5 Retrieval MCP tools.
The 20 curated Audit MCP tools.
Multi-tool recipes (Retrieval + Audit).
What we commit to keep stable, deprecation policy.
Last updated