featuresJuly 17, 2026· 4 min read

Give Claude and your agents a cache via MCP

The Crowkis binary doubles as an MCP server, so Claude Desktop, Claude Code, and any MCP-capable agent can check the cache before calling the model and store what they compute.

The Model Context Protocol flips the integration: instead of your code calling the cache around the model, the assistant itself holds cache operations as tools. Crowkis speaks JSON-RPC 2.0 over stdio with zero new dependencies.

mcp config

{
  "mcpServers": {
    "crowkis": { "command": "crowkis", "args": ["mcp"] }
  }
}

It registers tools the assistant can invoke mid-conversation, cache_get, cache_set, reuse, think, and stats, plus resources for live stats and the dashboard. A hit returns the answer with its confidence; a miss returns a clean signal the assistant can act on rather than hallucinate around.

In plain words: Your AI assistant gets a memory that lives on your machine. Questions it has already answered stop costing you money.

Crucially, MCP writes get no trust shortcut: an assistant's store request walks the same anti-poisoning pipeline as every other client, tracked as its own source. An agent can't poison the cache any more easily than a human can. The result is agents that remember as a behaviour, one config block, and the binary was already running.