Evo AI is an open-source platform for creating and managing AI agents, designed for developers and researchers building complex AI systems. It offers a flexible framework for integrating various language models, managing agent interactions, and orchestrating multi-agent workflows, aiming to simplify the development of sophisticated AI applications.

How It Works

Evo AI utilizes a modular architecture built with FastAPI, SQLAlchemy, and PostgreSQL. It supports diverse agent types including LLM, A2A (Agent-to-Agent protocol), Sequential, Parallel, Loop, and Workflow agents. Workflows can be visually designed using ReactFlow and executed via LangGraph. The platform emphasizes interoperability through the A2A protocol and provides secure API key management with Fernet encryption.

Quick Start & Requirements

• Install: git clone https://github.com/EvolutionAPI/evo-ai.git, cd evo-ai, make venv, source venv/bin/activate, pip install -e .
• Prerequisites: Python 3.10+, PostgreSQL 13.0+, Redis 6.0+, Git, Make, SendGrid Account (or SMTP server).
• Setup: Requires configuring .env file with database, JWT, and email credentials. Database migrations and seeding are handled via make alembic-upgrade and make seed-all.
• Docs: Swagger UI, ReDoc.

Highlighted Details

• Supports Google's Agent Development Kit (ADK) and Agent 2 Agent (A2A) protocol for interoperability.
• Integrates with Langfuse for tracing and observability via OpenTelemetry.
• Enables visual workflow creation with ReactFlow and execution with LangGraph.
• Features secure, encrypted API key management and agent organization via folders.
• Offers JWT authentication with email verification and role-based access control.

Maintenance & Community

The project is actively maintained by EvolutionAPI. Contribution guidelines are provided, encouraging community involvement.

Licensing & Compatibility

Licensed under the Apache License 2.0. The use of the "Evolution API" name, logo, or trademark is restricted and requires explicit permission as per section 6 of the license.

Limitations & Caveats

The frontend is a separate repository (evo-ai-frontend) requiring separate setup for full platform functionality. The README does not specify performance benchmarks or detailed system requirements beyond core dependencies.