Coding agents & developer tools
Autonomous coding agents
4,327 agents · most starred first
Autonomous coding agents, page 44 of 44
Web console and diagnostic workbench for AI coding agents: Elysia + Preact on Bun.
The AlignDev CLI: the assistant's playbook, coding-agent delegation, and project inventory.
AI Factory: requirement-driven, multi-agent code generation. The factory CLI.
An autonomous coding agent for the terminal
A control tower and phone-friendly review board for teams of AI coding agents (OpenCode, Claude Code, Codex).
A modular, hackable terminal coding agent built for long-running tasks and local models.
中文优先的终端编码智能体:读代码、改代码、跑命令;支持 DeepSeek、Kimi、通义、GLM、Claude 等,兼容 Claude Code 配置
Run the Quebra-Galho (qg) coding agent from Python.
Autonomous coding agent CLI — writes, edits, and builds code autonomously until the project compiles without errors.
Local-first agentic software engineering interface
An AI software team that runs in your GitHub org: shape, design, build, review, secure and ship.
Developer Foundation Model: a local, privacy-first engineering agent for Apple Silicon.
Self-improving coding agent built on the NOOA framework
A minimal, Pythonic coding agent: streaming LLM client, async-generator agent loop, and terminal harness.
Open-source (AGPL-3.0), self-hosted AI coding agent on Cloudflare. Approve each task, run it in an isolated sandbox, review the diff, open a PR. Works with OpenCode, Claude Code, Codex, Cursor and Grok.
Concord Cloud gives teams one hosted workspace where every developer's AI coding agents message each other, claim work and hand off, so the team stops colliding.
HiraxCode - terminal AI agent for coding and ethical hacking (pentest)
A lightweight AI agent for the terminal that works with any LLM, free tiers included: files, commands, web, MCP
An autonomous coding agent for your terminal.
The (AI) Framework: turnkey, zero-config AI orchestration that wraps a coding-agent CLI (Claude Code) as a black box and takes you from an idea to a running app. Vite for AI.
Code Automaton — a terminal agentic coding tool with session recall.
The simplest coding agent — six tools, one loop
Zerg Client is the coding agent harness Zerg AI uses every day: terminal, desktop app or headless, a model per pane. Because it owns its screen layer, the agent can open and drive terminal panes (zmux) while you watch — ask for vim or htop and panes open beside it. Available models depend on your sign-in route and provider access; desktop beta is Mac with Apple chips only.
Devin is Cognition's autonomous software engineer: it takes an engineering task and works it end to end — reading the codebase, writing the change and opening the pull request — rather than completing code inline as an editor assistant does. It is the name that recurs most often in 2026 roundups of autonomous coding agents, and it is sold per seat at a developer-tool price point rather than per token. Catalogued from Cognition's own product page; MeshKore does not operate it.
Software built by communities and shipped by AI agents. Sign up free and post the first idea.
OxCode is an AI software engineer, not an autocomplete tool. It works through the full engineering loop: it explores your codebase, plans the work, edits files, runs approved commands, tests, debugs and verifies the result before handing back, one reviewable step at a time.
A self-hosted, open source (AGPL-3.0) lights-out software factory where AI coding agents do the work. It turns GitHub Issues into tagged releases: independent review repairs the code, the exact reviewed head merges, and the frozen SHA ships as a published tag. Bring your own model.