ChatPilot provides a web UI for LLM-powered agents, enabling conversational AI with features like Google search, RAG for file/URL context, and a code interpreter. It targets users who want to replicate advanced chatbot functionalities similar to Kimi Chat, supporting various LLM backends including OpenAI, Azure, and local Ollama models.

How It Works

Built on Agentica, ChatPilot orchestrates agent interactions with external tools. It leverages a FastAPI backend and a Svelte frontend for a decoupled architecture. Key features include web search via Serper/DuckDuckGo, URL parsing, a Python code interpreter (with E2B or local execution), and RAG for document Q&A. It supports multiple LLM integrations through Ollama, LiteLLM, and OpenAI APIs.

Quick Start & Requirements

• Docker: export OPENAI_API_KEY=sk-xxx; export OPENAI_BASE_URL=https://xxx/v1; docker run -it -e OPENAI_API_KEY=$WORKSPACE_BASE -e OPENAI_BASE_URL=$OPENAI_BASE_URL -e RAG_EMBEDDING_MODEL="text-embedding-ada-002" -p 8080:8080 --name chatpilot-$(date +%Y%m%d%H%M%S) shibing624/chatpilot:0.0.1
• Local: git clone https://github.com/shibing624/ChatPilot.git; cd ChatPilot; pip install -r requirements.txt; cp .env.example .env; bash start.sh
• Prerequisites: OpenAI API key, potentially other LLM API keys/endpoints, Python >= 3.10 (for local), Node.js >= 20.10 (for frontend rebuild).
• Demo: https://chat.mulanai.com

Highlighted Details

• Supports voice input/output and image generation.
• Includes user management, permissions, and chat history import/export.
• Offers RAG capabilities for file and URL-based context.
• Integrates with E2B for secure code execution environments.

Maintenance & Community

The project is maintained by Ming Xu (xuming624@qq.com). Community engagement is encouraged via GitHub issues and WeChat.

Licensing & Compatibility

Licensed under the Apache License 2.0, permitting commercial use. A link to ChatPilot and the license must be included in product documentation.

Limitations & Caveats

The project is described as "rough" and welcomes contributions. Specific performance benchmarks or detailed limitations are not provided in the README.