- Multi-Model Orchestration: Seamlessly combine multiple AI models in customizable workflows
- Provider Flexibility: Support for multiple AI providers including DeepSeek, Claude, Gemini, and more
- OpenAI Compatible: Drop-in replacement for OpenAI's API in existing applications
- Stream Support: Real-time streaming responses for better user experience
- Advanced Configuration: Fine-grained control over model parameters and system prompts
- Database Integration: SQLite-based configuration storage with Alembic migrations
- Web Management UI: Built-in interface for managing models and configurations
git clone https://github.com/sligter/DeepGemini.git
cd DeepGemini
uv synccp .env.example .envRequired environment variables:
ALLOW_API_KEY: Your API access keyALLOW_ORIGINS: Allowed CORS origins (comma-separated or "*")
uv run uvicorn app.main:app --host 0.0.0.0 --port 8000Visit http://localhost:8000/dashboard to access the web management interface.
DeepGemini supports various AI providers:
- DeepSeek: Advanced reasoning capabilities
- Claude: Refined text generation
- Gemini: Google's AI model
- Custom: Add your own provider integration
Each model can be configured with:
- API credentials
- Model parameters (temperature, top_p, etc.)
- System prompts
- Usage type (reasoning/execution/both)
Create custom workflows by combining models:
- Reasoning Step: Initial analysis and planning
- Execution Step: Final response generation
- Custom Steps: Add multiple steps as needed
{
"name": "advanced_workflow",
"steps": [
{
"model_id": 1,
"step_type": "reasoning",
"step_order": 1
},
{
"model_id": 2,
"step_type": "execution",
"step_order": 2
}
]
}- FastAPI: Modern web framework
- SQLAlchemy: Database ORM
- Alembic: Database migrations
- UV: Fast Python package installer
- aiohttp: Async HTTP client
- deepclaude
This project is licensed under the MIT License - see the LICENSE file for details.
Contributions are welcome! Please feel free to submit a Pull Request.
For questions and support, please open an issue on GitHub.