Autonomous SWE agents
self-directed agents that go from issue to PR
code-yeongyu/oh-my-openagent code-yeongyu/oh-my-openagent is the clear leader by adoption and breadth — it’s the default if you want a well‑adopted, multi-agent orchestration harness for end‑to‑end repo workflows. Pick a different project only if you need a narrow capability (security, self‑hosting, sandboxed PRs, CI/test loops, or experimental autonomy) that the leader doesn’t prioritize.
Which one matches your setup?
Answer any of the questions — the shortlist updates as you go. Recommendations come from the capability passports below, nothing else.
Comparison matrix
Axes are extracted from each project's docs by our review pipeline; the maturity score is computed from stars, growth and commit activity — not an opinion. Click a column to sort.
| Runs in | Models | Context | Cost to run | |||||
|---|---|---|---|---|---|---|---|---|
⭐ 68.7k +214/7d | CLICoding-agent pluginIDE | BYOKAnthropicGeminiOpenAIFixed provider | Whole-repo analysis | Basic filters | ●●●●● | Cloud APIs (your key) | ●●●●● | Requires your provider subscriptions / API keys |
⭐ 33.8k +524/7d | CLIIDECoding-agent plugin | AnthropicOpenAIGeminiFixed provider | Whole-repo analysis | Basic QA (visual diff) | ●●●●● | Cloud agent APIs | ●●●●● | Requires an AI coding agent — provider billing applies |
⭐ 27.0k +123/7d | PR botCICLIWeb app | Fixed provider | Diff + related files | Human acceptance gating | ●●●●● | Self-hostable | ●●●●● | Self-hosted — runs on your infrastructure |
⭐ 20.2k +57/7d | CLIIDE | BYOK | Diff + related files | Basic config filters | ●●●●● | Your API key | ●●●●● | Your API key (cloud LMs) |
⭐ 16.5k +43/7d | CLIWeb appCI | BYOK | Whole-repo analysis | Permission & validation | ●●●●● | BYO API key | ●●●●● | Requires your model API key (OpenRouter or other provider) |
⭐ 10.7k +34/7d | PR botWeb appCLI | BYOK | Whole-repo access | Basic filters & middleware | ●●●●● | Cloud APIs (your key) | ●●●●● | Your API key + cloud sandbox costs |
⭐ 4.6k +62/7d | Web appPR botCI | Fixed provider | Diff + related files | None mentioned | ●●●●● | Self-hostable | ●●●●● | Free to start; online service or self-hosted |
⭐ 11.4k +68/7d | CLI | Fixed provider | Diff only | No controls | ●●●●● | Third-party cloud | ●●●●● | Unclear; depends on deployment |
⭐ 5.3k +12/7d | Web appCLICI | Gemini | Diff + related files | Linting & checks | ●●●●● | Cloud APIs (your key) | ●●●●● | Cloudflare Workers paid plan + Google Gemini API key |
⭐ 4.8k +28/7d | CLICoding-agent plugin | BYOKOllama / local | Related files + context | Approval gates + validation | ●●●●● | Fully local available | ●●●●● | Your model provider costs (use your API key) |
⭐ 3.4k +50/7d | CLIIDECoding-agent plugin | OpenAIAnthropic | Whole-repo analysis | Verified completion & review | ●●●●● | Cloud APIs (BYO key) | ●●●●● | Your API key, model provider charges |
⭐ 1.3k +29/7d | CLI | BYOKOllama / local | Whole-repo analysis | Reviewed workflows | ●●●●● | Fully local capable | ●●●●● | Your API key or a local GGUF model |
Capability profiles
Six axes, 0–5 each. The shape tells you the strategy: a wide hexagon is a generalist, a spike is a specialist. Showing the 8 most established — the rest are in the full catalog.
Orchestrates multiple agent harnesses and parallel "discipline" agents (Team Mode) to run whole-repo autonomous workflows rather than a single-model single-agent approach.
Reconstructs a live website into a production-ready Next.js codebase by extracting design tokens, assets, and exact computed styles, then dispatching parallel builder agents to assemble the site.
Orchestrates isolated, autonomous implementation runs so teams manage work instead of supervising individual coding agents.
State-of-the-art open-source agent focused on autonomously fixing GitHub issues and offering a specialized EnIGMA mode for offensive cybersecurity, all governed by a single YAML.
DeepCode pairs multi-agent delegation and reusable SKILL playbooks with autonomous, test-driven loops and parallel workers so it can decompose large features, run them in isolation, and iterate until tests pass.
Provides pluggable isolated cloud sandboxes and subagent orchestration out of the box for safe, parallelized repo edits and automatic PR creation.
Provides an enterprise-grade, end-to-end AI development platform that combines cloud development environments, AI task and model management, and requirement tracking for team workflows.
This repository is an archived/deprecated reference implementation and public issue history that points to an actively rebuilt service at humanlayer.com.
All repositories (46)
Clone any website with one command using AI coding agents
Symphony turns project work into isolated, autonomous implementation runs, allowing teams to manage work inste…
SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice. It can also be empl…
"DeepCode: Open Agentic Coding (Agent Harness & Loop Engineering & Multi-Agent Orchestration)"
The best way to get AI coding agents to solve hard problems in complex codebases.
An open-source vibe coding platform that helps you build your own vibe-coding platform, built entirely on Clou…
AI agent framework for plan-first development workflows with approval-based execution. Multi-language support…
The one and only agent harness for complex codebases. Project memory, planning, execution, and verified comple…
Ship your code, on autopilot. An open source agent that lives on your machines 24/7 and keeps your apps runnin…
Mention any ACP coding agent from Slack, GitHub, GitLab, Linear, or Lark. OpenTag runs Claude Code, Codex, Cur…
Your AI forgets. This remembers. Spec-driven coding harness for vibecoders, product owners, CEOs and real buil…
OpenAlpha_Evolve is an open-source Python framework inspired by the groundbreaking research on autonomous codi…
Autonomous software engineering fleet of AI agents for production-grade PRs on AgentField: plan, code, test, a…
Autoprompt is a coding-agent skill that cuts failures by 45% on agentic coding tasks.
Spec-driven development workflow for AI coding agents: architecture-first planning, task decomposition, GitHub…
Official AHE code — Agentic Harness Engineering: observability-driven automatic evolution of coding-agent harn…
Repeatable agents-plus-code workflows, packaged as one skill, stamped into any repo. Deterministic Python owns…
Claude Code plugin for Elixir/Phoenix/LiveView — 26 specialist agents, Iron Laws enforcement, and Tidewave MCP…
Bring your own agent and build a self-improving agentic system. Automatically mine failures, optimize the agen…
A Claude Code plugin that interviews you, designs the whole architecture, and writes a self-contained blueprin…
Codex/Claude Code/Cursor的开源增强版，专注一句话实现复杂长程任务（教育、编程、办公、生活、娱乐、游戏）。部署在你自己的服务器上，团队用浏览器打开就能编程——包括手机。CLI & Web UI 双入…
The agent skills I actually use to build software with coding agents. The PIV loop, planning, worktrees, and t…
A Boris-style agentic orchestrator TUI that supervises headless Claude Code agents through a gated software-de…
DeepAgent Code: AI coding agent with persistent memory and control plane
ATLAS: a senior-engineer layer for Claude Code. Explore with wireframes & prototypes, clarify the essentials,…
Local-LLM-first agentic coding assistant, with everything you need out of the box.
Broccoli turns Linear tickets into shipped PRs — powered by Claude and Codex, running on your own Google Cloud…
From ticket to reviewed pull request. Free and open-source, on your machine.
A local-first, open-source coding agent for your desktop. Bring your own LLM key; your code stays on your mach…
Universal, model-agnostic operating harness for AI agents (Claude, Codex, Gemini, …) — a lean core + work-type…
A local-first, cross-platform Electron desktop workspace for Pi Coding Agent, with sessions, project files, br…
Open-source OpenAI Codex alternative | Multi-agent desktop GUI for OpenCode | manage coding sessions, visualiz…
A plugin harness for Claude Code and Codex, built on bounded-autonomy architecture
Terminal-first, knowledge-grounded multi-agent software delivery pipeline: scope requirements, implement chang…
Terminal-first, knowledge-grounded multi-agent software delivery pipeline: scope requirements, implement chang…
The token-efficient agentic coding workbench. Built for a future where every token counts — it optimizes token…
Finish-First Autonomous Agent Loop for Claude Opus 4.7 – 2026 Edition
The agent harness that makes small local models finish hard tasks — and prove it. Done is enforced by the engi…
Hosted alternatives
If running your own reviewer is more ops than you want, these managed services cover the same job.
Cognition's hosted autonomous engineer: assign it a ticket and it plans, edits and opens a PR in its own cloud sandbox. Closest managed analogue to running SWE-agent yourself.
Try Devin →Async agent that clones your repo into a VM, works the issue and returns a diff; free tier for GitHub, no self-hosting.
Try Google Jules →Managed hosting from the OpenHands team — the same open-source agent, run for you with GitHub integration instead of your own infra.
Try OpenHands Cloud →