# Mem0 Docker with Qdrant

Dockerized memory service using mem0 with Qdrant vector database for semantic memory storage.

## Quick Start

1. Copy the example environment file:
```bash
cp example.env .env
```

2. Edit `.env` with your configuration:
   - `MEM0_PORT`: Port for the mem0 API
   - `QDRANT_HOST`: Qdrant host (default: qdrant)
   - `QDRANT_PORT`: Qdrant port (default: 6333)
   - `EMBEDDING_URL`: llama.cpp embedding endpoint URL
   - `EMBEDDING_DIMS`: Embedding dimension size

3. Start the services:
```bash
docker-compose up -d
```

## API Endpoints

### Health Check
```
GET /health
```

### Add Memory
```
POST /add
Content-Type: application/json

{
  "message": "Memory text to store",
  "user_id": "default",
  "metadata": {}
}
```

### Search Memories
```
POST /search
Content-Type: application/json

{
  "query": "Search query",
  "user_id": "default",
  "limit": 5
}
```

### Get All Memories
```
GET /memories?user_id=default
```

### Delete Memory
```
DELETE /delete/{memory_id}
```

## Files

- `docker-compose.yml`: Docker Compose configuration
- `Dockerfile`: Container build instructions
- `main.py`: FastAPI memory API server
- `mem0_server.py`: Alternative HTTP server implementation
- `requirements.txt`: Python dependencies
- `example.env`: Environment variable template

## Configuration

The service requires an external llama.cpp embedding endpoint. Configure the `EMBEDDING_URL` to point to your embedding service.