Fast Agency

FastAgency ? Accelerating Multi-Agent AI Workflows from Prototype to Production

FastAgency is an open-source framework designed to streamline the deployment of multi-agent AI workflows, facilitating the transition from prototype to production environments. It offers a unified programming interface, enabling developers to deploy agentic workflows across various platforms with minimal code modifications.

Key Features:

Unified Programming Interface: Develop workflows once and deploy them across different user interfaces, including console-based applications via ConsoleUI and web-based applications using MesopUI, without the need for extensive code rewrites.

Seamless External API Integration: Easily incorporate external APIs into agent workflows by importing OpenAPI specifications and connecting them with a single line of code, enhancing the functionality of AI agents with real-time data and third-party services.

Command-Line Interface (CLI) for Orchestration: Manage and automate workflows efficiently through a robust CLI, supporting DevOps and deployment tasks, and enabling developers to orchestrate complex workflows directly from the terminal.

Tester Class for Continuous Integration: Utilize the Tester Class to write and execute tests for multi-agent workflows, ensuring reliability and robustness, and integrating seamlessly with continuous integration pipelines.

Supported Runtimes and Adapters:

Runtime Support: Currently supports the AG2 (formerly AutoGen) framework, with plans to add support for additional frameworks like CrewAI, providing flexibility in choosing the appropriate runtime for specific applications.

Network Adapters: Offers chainable network adapters, including REST API via FastAPI and NATS.io via FastStream, facilitating scalable and production-ready architectures for serving workflows across multiple machines and data centers.