Cheshire Cat AI

Production ready AI agent framework

Free

Description

Cheshire Cat AI provides a robust, open-source framework designed for developers to build sophisticated AI agents. It allows users to train agents on custom documents, including PDF, TXT, Markdown, JSON files, and web pages. The framework facilitates seamless integration with external APIs and applications, enabling agents to interact dynamically with the digital world.

Built with a microservice-first approach and fully dockerized for easy deployment, Cheshire Cat AI supports both commercial and open Large Language Models (LLMs) and embedders. Developers can extend the agent's capabilities through a straightforward plugin system using Python, adding custom hooks, tools (function calling), and forms for complex, goal-oriented dialogues. An admin panel is included for managing installations, plugins, memory, configurations, and users, streamlining the development and deployment process.

Key Features

Document Training: Train agents by uploading various document formats (pdf, txt, md, json, web pages).
API & App Integration: Connect agents easily to external APIs and applications.
Flexible Model Support: Utilize commercial or open-source Large Language Models (LLMs) and embedders.
Dockerized Deployment: Packaged as a single container for easy integration and deployment, with live reload.
Extensible Plugin System: Enhance agent functionality with community or custom-built plugins.
Smart Dialogue Management: Implement advanced conversational skills with hooks, tools (function calling), and forms.
Developer-Focused Tools: Includes an admin panel, extensive APIs (HTTP/WebSocket), and community-built clients.
Microservice Architecture: Designed for easy integration as a conversational layer into existing software.