EvoAgentX

EvoAgentX is an open-source framework for building, evaluating, and evolving LLM-based agents.

EvoAgentX is an open-source framework for building, evaluating, and evolving LLM-based agents and agentic workflows. It enables developers and researchers to move beyond static prompt chaining through automated, modular, and goal-driven construction. The framework features built-in evaluation, a self-evolution engine, plug-and-play model compatibility with OpenAI, Qwen, Claude, DeepSeek, and Kimi via LiteLLM, SiliconFlow, or OpenRouter, and a comprehensive memory module supporting both ephemeral and persistent memory. It also includes human-in-the-loop interactions and diverse built-in tools such as code interpreters, search tools, filesystem utilities, and databases. Designed for AI researchers, workflow engineers, and startup teams, EvoAgentX provides a complete environment to develop and optimize real-world AI agents.

EvoAgentX at a glance

Best for: Software engineer, Researcher
Runs on: Linux, macOS, Windows, API, self-hosted, open source
Pricing: Free
Released: April 2025
Category: AI agent tool
Website: github.com/ANative-Lab/EvoAgentX

Key features

Agent workflow autoconstruction
Built-in evaluation
Self-evolution engine
Plug-and-play model compatibility
Comprehensive built-in tools
Short-term and long-term memory module
Human-in-the-loop interactions

Use it for

Optimize agentic workflows using self-evolving algorithms
Supervise agents with human-in-the-loop checkpoints
Maintain context across interactions using memory modules
Automatically assemble multi-agent workflows from goals
Interact with real-world environments using built-in tools

Alternatives to EvoAgentX

OmniAgent — It evolves skills in real-time and dynamically hardens security during task execution.
Evolver — Reusable solutions and consensus validation improve reliability and consistency.
EvoScientist — It features a self-evolving memory graph that compounds skills across research runs.
AGiXT — Builds autonomous agents with integrated memory and web capabilities.
OpenEvolve — Automates algorithmic discovery and optimization without human guidance.
Flowise — Modular building blocks and visual tools to build and scale agentic systems fast.
CAMEL — Builds complex multi-agent societies for data generation and simulation.
Composio — Equips AI agents with tools, managed user auth, and sandboxes across 1,500+ apps
Harness2 — It recursively refines agent execution structures using parallel and sequential search.
FastGPT — It uses hybrid retrieval and auto-cleansed data to prevent hallucinations.
VoltAgent — Offers open-source flexibility with built-in supervisor agent orchestration and RAG.
GEAR — Organizes recursive self-improvement through evolutionary search and reproducible evaluation.

All EvoAgentX alternatives