A conversational chatbot API built with FastAPI, PostgreSQL, ChromaDB, and LM Studio for local LLM inference.

• 🤖 AI-Powered Chat: Connect to LM Studio for local LLM inference
• 👥 User Management: Create, read, update, delete users
• 💬 Conversation Management: Organize chats into conversations
• 📝 Message History: Store and retrieve conversation history
• 🔍 Vector Search: ChromaDB integration for RAG capabilities
• 🐳 Docker Ready: Full Docker Compose setup for easy deployment

botgpt/
├── app/
│   ├── main.py              # FastAPI application entry point
│   ├── core/                # Business logic (LLM client, embedder)
│   ├── db/                  # Database clients (Postgres, ChromaDB)
│   ├── models/              # SQLAlchemy ORM models
│   ├── routes/v1/           # API endpoints
│   ├── schemas/             # Pydantic schemas
│   ├── utils/               # Utilities (logging, security)
│   └── docs/                # Documentation
├── tests/                   # Test files
├── docker-compose.yml       # Docker services
├── Dockerfile               # Application container
└── requirements.txt         # Python dependencies

• Docker and Docker Compose
• LM Studio running on your host machine
• Python 3.13+ (for local development)

git clone https://github.com/maharanasarkar/bot-gpt.git
cd bot-gpt

1. Download and install LM Studio
2. Load a model (e.g., qwen/qwen3-4b)
3. Start the local server on port 1234

# Build and start all services
./docker.sh

# Or manually:
docker build -t botgpt:1.0.0 .
docker-compose up -d

• API Base URL: http://localhost:5000
• Swagger Docs: http://localhost:5000/docs
• ReDoc: http://localhost:5000/redoc

Configure these in docker-compose.yml or a .env file:

curl -X POST http://localhost:5000/v1/users/ \
  -H "Content-Type: application/json" \
  -d '{
    "email": "[email protected]",
    "password": "mypassword",
    "first_name": "John",
    "last_name": "Doe"
  }'

curl -X POST http://localhost:5000/v1/conversations/ \
  -H "Content-Type: application/json" \
  -d '{
    "user_id": "<USER_ID>",
    "title": "My First Chat"
  }'

curl -X POST http://localhost:5000/v1/messages/ \
  -H "Content-Type: application/json" \
  -d '{
    "user_id": "<USER_ID>",
    "conv_id": "<CONVERSATION_ID>",
    "message_sender": "user",
    "content": "Hello, how are you?"
  }'

Response:

{
  "user_message": {
    "id": "msg123",
    "content": "Hello, how are you?",
    "message_sender": "user",
    ...
  },
  "model_response": "Hello! I'm doing well. How can I assist you today?"
}

curl "http://localhost:5000/v1/messages/?conv_id=<CONVERSATION_ID>"

python -m venv .venv
source .venv/bin/activate  # Linux/Mac
# or
.venv\Scripts\activate     # Windows

pip install -r requirements.txt

# Start PostgreSQL and ChromaDB with Docker
docker-compose up -d postgres chromadb

# Run FastAPI locally
uvicorn app.main:app --reload --port 5000

pytest tests/

• API Reference - Complete API documentation
• Architecture - System design and structure

If you see "Client not initialized" errors:

1. Ensure LM Studio is running on your host machine
2. Verify the server is started on port 1234
3. Check that LLM_BASE_URL is set to http://host.docker.internal:1234/v1

1. Check PostgreSQL container is running: docker-compose ps
2. Verify DATABASE_URL format: postgresql+asyncpg://user:pass@postgres:5432/dbname

If ports are already in use, modify the port mappings in docker-compose.yml.

MIT License

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Open a Pull Request