EvoAgentX
EvoAgentX — a framework for building self-evolving ecosystems of AI agents.
A GEP-powered self-evolving engine for AI agents — auditable evolution with Genes, Capsules and Events.
evolver is an open-source ai agents project by EvoMap (EvoMap/evolver on GitHub). It has 9,138 GitHub stars and 848 forks, is released under the GPL-3.0 license, and was last updated on GitHub on 30 September 2026.
Pick the git repo and agent whose logs and prompts you want to evolve, and decide whether evolution runs in review mode or as a background loop.
Clone EvoMap/evolver from GitHub (or use the install path in the run steps below) and skim its README for the parts relevant to your goal.
Install the CLI globally from npm; Node.js 18+ and git are prerequisites, and no clone is needed unless you are contributing.
Run evolver once in the repo for a single evolution run, review the selected Gene and GEP prompt, then decide whether to enable the continuous loop.
Improvements are captured as Genes, Capsules and Events — auditable, reusable assets rather than one-off prompt edits.
Run one evolution pass on demand, a review mode that waits for human confirmation, or a continuous background loop.
Evolver uses git for rollback, blast-radius calculation and solidifying accepted changes, and refuses to run outside a git repo.
Optionally register with the EvoMap evolution network to connect your node to the wider ecosystem.
Optional integration points wire the evolved output back into your agent runtime.
The README is explicit about what Evolver does not do — it does not auto-edit your source code or execute arbitrary shell commands outside its security model.
Get the source from the official GitHub repository for evolver.
git clone https://github.com/EvoMap/evolver.git
cd evolverInstall Evolver globally from npm as the README's 30-second quick start shows.
npm install -g @evomap/evolverRun evolver inside any git repo to scan logs, select a Gene and output a GEP prompt; add a .env in that working directory if your setup needs one.
evolver # run inside your git repo — see README for review and loop modesEvoMap, through the EvoMap/evolver repository, with the network and wiki at evomap.ai.
Yes — the project is open source under the GPL-3.0 license as recorded for this listing. See the license terms for any conditions on redistribution.
Evolver drives evolution using model access you configure in a .env in the working directory; the tool itself is open source, and any model cost is yours.
Manual tuning is fast but unauditable and easy to lose. Evolver formalises the same loop into Genes, Capsules and Events with git rollback and review modes — more process, in exchange for a history you can inspect and revert.
Clone the repository (or use the documented installer), follow the install and configure steps on this page, and run the smallest real task first — run evolver once in the repo for a single evolution run, review the selected Gene and GEP prompt, then decide whether to enable the continuous loop.
EvoAgentX — a framework for building self-evolving ecosystems of AI agents.
Reef — infrastructure for continually self-improving agents: weight-training and harness-evolving deployments.
Self-evolving memory OS for LLMs and agents: persistent graph memory, hybrid retrieval, skill reuse.
Your self-hostable AI second brain for answers, agents, and automations.
SimWorld: an Unreal Engine 5 simulator for evaluating LLM/VLM agents in physical and social worlds.
eBPF-powered information-flow enforcement for safer AI agents.
The most-starred other AI Agents projects in the LibreAI catalogue — the realistic shortlist if evolver is not the right fit.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Free and open — grab the source, star the repo, and start creating.