autonomous AI agent orchestration / unattended coding agents
An autonomous AI-agent framework/harness that runs unattended in CI (GitHub Actions), driving coding CLIs like Claude Code, Codex and Grok with self-healing skills and MCP tools.
Solves the same problem with a different architecture or at a different layer. Expect to rewrite the integration.
Pick it when you want a general agent platform with hosted UI and marketplace instead of CI-scheduled runs.
AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.
Pick it for a sandboxed Docker runtime with web UI and its own agent loop, at the cost of much heavier infra than a GitHub Actions job.
🙌 OpenHands: AI-Driven Development
TypeScript RLM agent for long-running coding; pick it if you want in-process self-improvement rather than scheduled CI invocations.
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
Go terminal agent with its own sandboxed diff workflow for large repos; pick it for interactive work, but note the last push was 12 months ago.
Open source AI coding agent. Designed for large projects and real world tasks.
Go self-hosted assistant with computer-use and browser-use; pick it when agents must drive a desktop rather than a CI runner.
⚡️next-generation personal AI assistant powered by LLM, RAG and agent loops, supporting computer-use, browser-use and coding agent, demo: https://demo.openagentai.org
Python framework for low-level agent internals; pick it for fine-grained control, accepting GPL-3.0 and a small 847-star community.
Extensible AGI Framework
Python self-improving agent; pick it when you want the self-improvement loop as an embeddable library rather than a CI harness.
self-improving agent
Vision-grounded computer-use agents with self-written skills; pick it for desktop automation, not repo automation.
Autonomous self-evolving agents. Vision-grounded layered memory and self-written skills for LLM agents that operate your computer.
TypeScript self-hosted multi-agent runtime with scheduling and 23+ providers; pick it to run swarms on your own box instead of GitHub-hosted runners.
Open-source self-hosted AI agent runtime and multi-agent framework for autonomous agent swarms. Agent memory, MCP tools, schedules, delegation, and 23+ LLM providers (Claude, GPT, Gemini, OpenRouter, Ollama). A practical Claude Code and LangChain alternative.
JavaScript local-first agent OS; pick it to build and debug agents on a workstation rather than in scheduled cloud jobs.
The local-first operating system for building, running, and improving AI agents, workflows, and autonomous goals.
Python multi-agent harness with production observability; pick it for a long-lived server deployment instead of per-run CI agents.
Multi-Agent Harness for Production AI
Python agent that accumulates skills over time; pick it for a persistent personal agent rather than unattended scheduled runs.
The agent that grows with you
Python terminal agent with local tools; pick it for an interactive local agent instead of unattended CI execution.
Your agent in your terminal, equipped with local tools: writes code, uses the terminal, browses the web. Make your own persistent autonomous agent on top!
A bash loop over Claude Code with exit detection; pick it when one script beats a full skills/MCP framework.
Autonomous AI development loop for Claude Code with intelligent exit detection
TypeScript harness builder that structures coding work around a knowledge base; pick it for deterministic, repeatable coding workflows.
The first open-source harness builder for AI coding. Make AI coding deterministic and repeatable.
Rust single-binary orchestrator of the Ralph loop; pick it when you want compiled scheduling instead of shell scripts.
An improved implementation of the Ralph Wiggum technique for autonomous AI agent orchestration
An earlier or more limited way to do the job. Still the right call when you need something small, proven, or CPU-only.
Markdown-only Claude Code skills implementing a research loop; pick it for ML research automation only, with no general devops agent support.
ARIS ⚔️ (Auto-Research-In-Sleep) — Lightweight Markdown-only skills for autonomous ML research: cross-model review loops, idea discovery, and experiment automation. No framework, no lock-in — works with Claude Code, Codex, OpenClaw, or any LLM agent.
A single goal-directed iteration skill for Claude Code; pick it for a minimal, framework-free loop with no MCP or multi-provider handling.
Claude Autoresearch Skill — Autonomous goal-directed iteration for Claude Code. Inspired by Karpathy's autoresearch. Modify → Verify → Keep/Discard → Repeat forever.
Autonomous loop restricted to Godot/Bevy/Babylon game projects; pick it only for game-dev pipelines.
Autonomous game development for Godot, Bevy, and Babylon.js with Claude Code and Codex
An experiment-loop extension living inside the pi agent; pick it for narrow experiment automation inside pi only.
Autonomous experiment loop extension for pi
These projects were analysed and named aeon among their alternatives. The relationship is not symmetric — how aeon rates them is a separate judgement, made when aeon is analysed in its own right.
calls aeon “Pick it when you want unattended, self-healing runs on CI (GitHub Actions) driving multiple external agents instead of an interactive local harness.”
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
calls aeon “Choose when you want unattended GitHub Actions-based autonomous agents and can accept a shell-driven runtime.”
BabyAGI: an Autonomous and Self-Improving agent, or BASI
calls aeon “Choose for unattended GitHub Actions orchestration of external coding agents rather than Automata's self-coding loop.”
Automata: A self-coding agent
calls aeon “Choose it to run unattended agents on GitHub Actions if you prefer a shell-based framework over a desktop app.”
The local-first operating system for building, running, and improving AI agents, workflows, and autonomous goals.