Langchain-Chatchat

Stale

GitHub Python Apache-2.0

Description

A local knowledge base RAG and Agent application platform built on Langchain with support for ChatGLM, Qwen, Llama and other LLMs, offering conversation, knowledge base management, and agent capabilities.

Key Features

Local knowledge base QA — Document loading, text splitting, vectorization and semantic retrieval via Langchain with BM25+KNN hybrid search
Multi-framework model inference — Supports Xinference, Ollama, FastChat, LocalAI for GLM-4, Qwen2, Llama3 and more
Agent tool calling — Optimized for ChatGLM3/Qwen with auto tool selection, single tool parameter parsing, and multimodal image chat
Database conversation — Natural language SQL queries against MySQL, PostgreSQL and other relational databases
WebUI multi-session management — Streamlit-based web interface with multi-session, custom system prompts, ARXIV and Wolfram queries
Full offline private deployment — 100% open-source models and vector databases via Ollama/Xinference local deployment

Use Cases

💡 Enterprise internal knowledge base — Upload company documents for private QA on policies, product manuals, and technical specs

💡 Offline data analysis — Query databases and generate analysis reports via local models without internet access

💡 Multimodal document QA — Upload images and documents for visual QA and document understanding with Qwen-VL

💡 Academic literature retrieval — ARXIV and Wolfram integration for paper search and data computation

💡 Customized agent services — Build domain-specific agents for customer service, tech support, and other business scenarios

Strengths & Limitations

✅ Strengths

• High community interest (38.6k stars)
• Permissive open-source license (Apache-2.0)
• Established track record (3 years in production)
• Responsive to issues, low backlog

⚠️ Limitations

• No updates in over 9 months

Quick Start

```bash
pip install langchain-chatchat
chatchat init
# Edit configs/model_config.py to configure models
chatchat kb -n my_knowledge  # Create knowledge base
chatchat webui  # Start WebUI
# Visit http://localhost:8501
```

Visit GitHub

Related Projects

Onyx

31.9k · Python

Active A

Open source AI platform with enterprise-grade AI chat, advanced RAG and AI search capabilities that works with every LLM.

ragpythonllm +2

· Agentic RAG — Hybrid index + AI Agent retrieval for best-in-class search and answer quality
· Deep Research — Multi-step research flow generating comprehensive reports, top of leaderboard
· Custom Agents — Build AI Agents with unique instructions, knowledge bases, and action capabilities

View full details → View on GitHub

RAGapp

4.4k · TypeScript

Stale C

The easiest way to use Agentic RAG in any enterprise. Provides out-of-the-box retrieval-augmented generation capabilities with Docker-based deployment for simplified enterprise RAG application building and management.

ragagentpython +2

· Docker one-command deployment with Admin UI for visual RAG configuration
· Agentic RAG pipeline built on LlamaIndex with OpenAI, Gemini, and Ollama support
· Pre-built Chat UI and REST API endpoints ready for integration

View on GitHub

MaxKB

22.7k · Python

Active A+

MaxKB is an open-source knowledge base Q&A and agent building platform powered by LLMs, with vector retrieval, workflow orchestration, and multi-model support out of the box.

knowledge-baseragchatbot +3

· RAG Pipeline — Document upload and online crawling with automatic splitting and vectorization to reduce hallucinations
· Agentic Workflow — Powerful workflow engine, function library and MCP tool-use for complex business scenarios
· Zero-code Integration — Seamlessly integrate into third-party business systems for smart Q&A capabilities

View full details → View on GitHub

MemAgent

1.1k · Python

Stale B

A MemAgent framework that can extrapolate to 3.5M context tokens, along with a training framework for RL training of any agent workflow.

memoryagentrag +2

· Ultra-Long Context Processing: Extrapolate from 8K training context to 3.5M tokens with performance loss under 5%
· Reinforcement Learning Driven: Trained with RLVR (Reinforcement Learning from Verifiable Rewards), extends DAPO algorithm for end-to-end multi-turn conversation optimization
· Linear Time Complexity: Breaks through computational bottlenecks in long-text processing with linear resource scaling

View on GitHub

RAG文档分块Chunking

Document Chunking Strategies: Recursive vs Semantic vs Agentic in Practice

RAG upstream caps downstream performance. A systematic comparison of four chunking strategies -- fixed-size, recursive, semantic, and agentic -- with quantitative selection guidance, Chonkie / LlamaIndex / LangChain tooling comparison, and offline evaluation methods.