An extensible AI agent microservice for building, testing, and deploying conversational AI assistants.
Build and experiment with custom AI agents using Python, function calling, and Model Context Protocol.
Cheshire Cat AI is an open-source Python microservice framework designed for building and publishing conversational AI agents. Insurance teams can use it to prototype custom virtual assistants and experiment with LLM function calling. It is best suited for research, learning, and early-stage agentic workflow prototyping.
The tool runs an AI agent server complete with a built-in web UI and OpenAPI REST playground on port 1865. It supports function calling, agentic engineering, and Model Context Protocol (MCP) client features out of the box. You can extend its logic using custom Python modules and connect it to your own data sources.
It suits insurance innovation teams, developers, and researchers building proof-of-concept AI assistants for policy or claims queries. It does not suit teams looking for immediate enterprise production deployments, as Version 2 is currently an unstable alpha release.
Access an interactive web user interface on port 1865 to test conversations with your AI agent.
Interact with agent API endpoints using the automatically generated Swagger documentation on port 1865/docs.
Connect your conversational agent to external context providers and tools through built-in MCP client features.
Extend agent capabilities by linking custom Python functions directly into LLM decision-making workflows.
Cheshire Cat AI is free under the GPL-3.0 license. There is no hosted paid version mentioned in the repository, so your only expenses are your self-hosting infrastructure and LLM API usage costs.
896 open remote AI-training roles, updated November 16, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
No, Version 2 is currently an unstable alpha release intended for research and educational experiments, not production workloads.
You can connect external tools and data sources using custom Python functions or built-in Model Context Protocol integrations.
See Zekai first in Google