core

Cheshire Cat AI is a Python-based framework for building custom AI agents, packaged as a fully dockerized microservice that can be launched with a single `docker run` command. It connects to language models through LangChain, meaning developers can point it at Claude or any other supported LLM, and exposes the agent via WebSocket chat and a REST API, making it straightforward to embed a conversational layer into existing applications. The framework includes built-in RAG using Qdrant for vector search, a plugin system built around hooks (event callbacks), function-calling tools, and structured conversational forms for goal-oriented dialogues such as order flows. It also acts as an MCP client, supporting the Model Context Protocol for tool integration. An admin panel handles multiuser access with granular permissions compatible with any identity provider. Developers building chatbots, internal assistants, or API-driven conversational products are the primary audience, and the plugin architecture lets them extend behavior without modifying core code.

What people ask about core