autonomous coding agents / self-improving agents
A Python framework for an autonomous self-coding agent that uses LLMs to write and evolve its own code toward AGI.
Solves the same problem with a different architecture or at a different layer. Expect to rewrite the integration.
Choose for a broader, actively maintained general-purpose autonomous agent platform rather than a self-coding framework.
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.
Choose for large-project codebase edits via a Go CLI, not self-modifying agent behavior.
Open source AI coding agent. Designed for large projects and real world tasks.
Choose for a general personal assistant with computer/browser/coding agents if you do not need self-modification.
⚡️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
Choose for a mature software-engineering agent that operates a full dev environment, at the cost of a larger platform.
🙌 OpenHands: AI-Driven Development
Choose for unattended GitHub Actions orchestration of external coding agents rather than Automata's self-coding loop.
The most autonomous AI agent framework: runs unattended on GitHub Actions, self-healing skills, drives Claude Code, Grok, Codex & more. No approval loops. Configure once, forget forever.
Choose for computer-use agents that self-write skills via vision-grounded memory instead of code-focused self-programming.
Autonomous self-evolving agents. Vision-grounded layered memory and self-written skills for LLM agents that operate your computer.
Choose for a multi-agent runtime with memory, MCP tools and scheduling rather than a self-coding agent.
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.
An earlier or more limited way to do the job. Still the right call when you need something small, proven, or CPU-only.
Choose for a minimal BabyAGI-style self-improving loop if you can accept an unmaintained codebase.
BabyAGI: an Autonomous and Self-Improving agent, or BASI
Choose for a tiny OpenAI-API autonomous-task script, but it is unmaintained and lacks self-coding.
MiniAGI is a simple general-purpose AI agent based on the OpenAI API.
Choose for a data-centric self-evolving language-agent framework if a two-year-old codebase is acceptable.
An Open-source Framework for Data-centric, Self-evolving Autonomous Language Agents
Choose for automating ML engineering experiments specifically, not general self-coding.
AIDE: an LLM agent for machine learning engineering - the research Weco grew out of. Referenced in OpenAI MLE-bench.
Does the same job, but ships as an app. Useful to a person, not swappable into a codebase.
Choose as an end-user CLI agent for terminal/browser/editor automation, not as an embeddable library.
✨ Fully autonomous AI Agent that can perform complicated tasks and projects using terminal, browser, and editor.
These projects were analysed and named automata among their alternatives. The relationship is not symmetric — how automata rates them is a separate judgement, made when automata is analysed in its own right.
calls automata “Choose only for self-coding agent experiments in Python; archived and much narrower than general goal agents.”
🤖 Assemble, configure, and deploy autonomous AI Agents in your browser.
calls automata “narrower self-coding agent framework, archived and unmaintained; choose only if you need a self-coding agent and accept archived code”
An Open-source Framework for Data-centric, Self-evolving Autonomous Language Agents
calls automata “Choose if you specifically need a self-coding agent and can accept an archived, stale Python codebase.”
BabyAGI: an Autonomous and Self-Improving agent, or BASI