One API for All AI Services - Build, orchestrate, and deploy intelligent AI agents with ease.
GTWY AI is an open-source unified AI middleware that simplifies AI integration across multiple providers. Create powerful AI agents, orchestrate multi-agent workflows, and connect to 5000+ apps - all through a single, consistent API.
- 🤖 Create AI Agents: Design custom AI agents with specific capabilities and personas
- 🔄 Orchestrate Multi-Agent Systems: Build complex workflows where agents collaborate and transfer tasks
- 💬 Deploy as API or Chatbot: Use your agents through REST APIs or integrate as conversational interfaces
- 📚 RAG as a Service: Implement retrieval-augmented generation with built-in document processing and vector search
- 🔌 5000+ App Connectors: Integrate with thousands of external services and tools through agent connections
For detailed architecture and flow documentation, see:
- Chat Completion Flow - Comprehensive API flow from request to response
- OpenAI: GPT models, embeddings, image generation, batch processing
- Anthropic: Claude models with vision support, batch processing
- Google Gemini: Text, image, and video processing capabilities
- Groq: High-performance inference with batch support
- Mistral AI: Advanced language models with batch processing
- OpenRouter: Access to multiple AI providers through a unified interface
- Grok: xAI's language models
- Custom AI/ML: Support for custom model endpoints
- Multi-Agent Orchestration: Configure and manage multiple AI agents with automatic tool generation and transfer capabilities
- Retrieval-Augmented Generation (RAG): Document chunking, vector search, and context-aware responses
- Function/Tool Calling: Built-in and custom tools with parallel execution support
- Batch Processing: Async batch job creation and monitoring across multiple providers
- Memory Management: Conversation history with configurable context windows
- Guardrails: Content filtering and validation
- Web Search Integration: Built-in web search tools with domain filtering
- Image Processing: Multi-modal support for image analysis and generation
- Pre-built Prompts: Reusable prompt templates and configurations
# Clone the repo
git clone https://github.com/Walkover-Web-Solution/AI-middleware-python.git
cd AI-middleware-python
# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate # Linux/Mac
venv\Scripts\activate # Windows
# Install dependencies
pip install -r req.txt
# Run the FastAPI server
uvicorn app.main:app --reload