🤖Chat_Agent

This is a great repo for beginners who are trying to practice LangChain and the repo is also my first project after learning LangChain

😊What's the project

An AI knowledge base Q&A assistant built on RAG and tool calling,supporting local document retrieval and MCP service web search.User can ask question in natural language;the system intelligently matches data sources to generate accurate answers,and features multi-turn diaogue context memory capability.

❓Target Audience

The beginners who want to learn LangChain through hands-on project
Individuals with a keen interest in cutting-edge AI technologies
Professionals planning a career transition

📁File structure

CHAT_AGENT
|
|---api/          chat & upload Router
|---model/        AI agent model
|---app/          back_end
|---util/         tool function
|---MCP/          integrated MCP service
|---tools/        Tool packaging
|---tests/        pytest test suite
|---.github/      GitHub Actions CI/CD
|---Dockerfile            Docker image build
|---docker-compose.yml    one-command orchestration
|---.dockerignore         Docker ignore rules

🤔What I learned

Learn about LangChain and FastAPI.
More structured project structure.
Not only API calls,but also tools encapsulation and MCP services.

🔭Quick start

Visit:https://chatagent-production-3489.up.railway.app/

⚙️How to run

🐳 Run with Docker (Recommended)

# Pull from GitHub Container Registry
docker pull ghcr.io/elysianx138/chat_agent:latest

# Run
docker run -p 8000:8000 --env-file .env ghcr.io/elysianx138/chat_agent:latest

🐳 Run with Docker Compose

docker compose up -d

🖥️ Run Locally (without Docker)

Step 1

Install requirements.txt

pip install -r requirements.txt

Step 2

Fill in the necessary API

copy .env.example .env

AI_MODEL=YOUR_MODEL
BASE_URL=YOUR_URL
API_KEY=YOUR_API_KEY
AI_EMBEDDING_MODEL=YOUR_AI_EMBEDDING_MODEL
SEARCH_API=YOUR_SEARCH_API

Step 3

Contact your knowledge base

 UPLOAD_DIR = Path(os.getenv("UPLOAD_DIR","uploads"))

Step 4

RUN

python -m app.main

Step 5

Uvicorn running on http://127.0.0.1:8000

Step 6

Input http://127.0.0.1:8000/docs#/default/chat_chat_post

👀Preview

☀ You can chat with this AI daily

🔍 Search with AI!

📚 Know your knowledge base

😄Q & A

Q:Why is there nothing afer opening the correct address? A:Because we don't have a front-end,if you want to try,need to enter:http://127.0.0.1:8000/docs#/default/chat_chat_post

Q:Why do I keep reporting errors when I use it? A:Try to check your API(AI model,embedding model,etc.) or check your uploadsfile whether the content of your file uses Mkdown?

Q:There is still a problem A:Welcome to submit Issues or contact me

😟Have any questions

submit Issues Contact the author

📃LICENSE

MIT License - See LICENSE fileLICENSE

elysianx138/Chat_Agent

README