GenAI Multi-Agent Orchestration Framework
A full-stack, production-ready framework for building, testing, and deploying sophisticated AI systems where multiple agents collaborate to solve complex problems.
Features at a Glance
🧩 Modular Agent Architecture
Plug-and-play structure for easily adding new, specialized AI agents.
🧠 LLM-Powered Orchestrator
A central routing agent that understands natural language and delegates tasks.
🔄 Multi-Agent Workflows
Execute complex, multi-step tasks by chaining agents together in sequence.
💾 Persistent Shared Memory
File-backed memory allows agents to share context and state across runs.
🌐 Triple Interface
Interact via a REST API (FastAPI), an interactive CLI (Typer), or a user-friendly Web UI (Streamlit).
🧪 YAML-Based Evaluation
A simple but effective testing suite to run predefined test cases and validate outputs.
Interactive System Architecture
The framework is designed with a clear separation of concerns. Hover over any component to learn more about its role in the system.
User Interfaces
Core System
Specialist Agents
Agent Explorer
Discover the specialist agents available in the framework. Click on any agent to see its description and required inputs. Use the search to filter agents.
Usage Guide
Interact with the system through its three primary interfaces. Use the tabs below to explore code examples for the API, CLI, and Web UI.
Use `curl` or any API client to interact with the backend. The server must be running.
Future Roadmap
Enhanced Logging & Trace Viewer
Implement a robust logging system and a UI to visualize the step-by-step execution of agent workflows.
Deployment Packaging
Create Dockerfiles and deployment scripts for easy hosting on platforms like Hugging Face Spaces or cloud services.
Vector-Based Memory
Upgrade the memory system to use a vector database for long-term, semantic memory retrieval.
Add More Agents
Continuously expand the system's capabilities by adding new specialist agents for different domains.