Awesome AI Dev Tools
Curated ranking of AI-powered developer tools, sorted by GitHub stars. Updated daily by GitHub Actions.
📊 Interactive dashboard with star-history charts →
Last updated: 2026-10-03 04:21 UTC · Tracked: 42 repos Top 7-day mover: deepseek-ai/deepseek-harness (+6290 stars)
| # | Repo | Stars | Δ7d | Language | Last push | Description |
|---|---|---|---|---|---|---|
| 1 | deepseek-ai/deepseek-harness | 242.4k | +6290 | TypeScript | 2026-10-03 | DeepSeek's official plugin-based agent harness (dsh), run through npx with a web UI |
| 2 | anomalyco/opencode | 211.5k | +1436 | TypeScript | 2026-10-03 | Open-source coding agent with a terminal UI and desktop app that works with any model provider |
| 3 | anthropics/claude-code | 149.0k | +879 | TypeScript | 2026-10-03 | Anthropic's terminal coding agent that reads your codebase, edits files, runs commands, and handles git workflows |
| 4 | openai/codex | 127.7k | +1150 | Rust | 2026-10-03 | OpenAI's coding agent that runs locally in your terminal, with IDE extension and desktop app options |
| 5 | earendil-works/pi | 111.8k | +2429 | TypeScript | 2026-10-02 | Toolkit with a unified multi-provider LLM API, an agent runtime, a terminal UI, and an extensible coding agent CLI |
| 6 | google-gemini/gemini-cli | 107.2k | +51 | TypeScript | 2026-10-03 | Google's open-source terminal agent for Gemini models, with built-in search, file and shell tools, and MCP |
| 7 | zed-industries/zed | 91.2k | +315 | Rust | 2026-10-03 | High-performance multiplayer code editor with a built-in agent panel that also hosts external agents over ACP |
| 8 | OpenHands/OpenHands | 89.8k | +656 | TypeScript | 2026-10-02 | Self-hosted control center that runs OpenHands or other ACP agents locally, in Docker, or in the cloud |
| 9 | stablyai/orca | 83.9k | +5536 | TypeScript | 2026-10-03 | Desktop agent development environment that runs Codex, Claude Code, OpenCode, or Pi side by side, each in its own git worktree |
| 10 | cline/cline | 69.7k | +408 | TypeScript | 2026-10-03 | Open-source coding agent for your IDE, terminal, or desktop, with a headless CLI mode for CI |
| 11 | openinterpreter/openinterpreter | 68.5k | +52 | Rust | 2026-10-02 | Terminal coding agent written in Rust, tuned for low-cost open-weight models |
| 12 | warpdotdev/warp | 65.3k | +184 | Rust | 2026-10-03 | Agentic development environment built on a terminal, with a built-in coding agent that also runs third-party CLI agents |
| 13 | AntonOsika/gpt-engineer | 55.1k | -20 | Python | 2025-05-14 | CLI that generates a codebase from a natural-language spec and iterates on feedback; archived upstream |
| 14 | aaif-goose/goose | 54.9k | +224 | Rust | 2026-10-02 | Extensible open-source agent in Rust with a desktop app, CLI, and API that works with any LLM and MCP extensions |
| 15 | Aider-AI/aider | 49.3k | +150 | Python | 2026-05-22 | Terminal pair-programming tool that maps your codebase and edits it with cloud or local LLMs |
| 16 | alibaba/open-code-review | 43.4k | +1996 | Go | 2026-10-01 | AI code review CLI that reads Git diffs and posts line-level comments, pairing deterministic checks with an LLM agent |
| 17 | Hmbown/Codewhale | 41.0k | 0 | Rust | 2026-10-02 | Open-source terminal coding agent in Rust that works with hosted or local models and delegates work to sub-agents |
| 18 | esengine/DeepSeek-Reasonix | 35.7k | +23 | Go | 2026-10-03 | DeepSeek-native coding agent as a single Go binary with terminal, desktop, browser, and ACP clients, built around prefix caching |
| 19 | can1357/oh-my-pi | 34.1k | +808 | TypeScript | 2026-10-03 | Terminal coding agent forked from Pi, with LSP and debugger (DAP) integration built in |
| 20 | TabbyML/tabby | 33.9k | -6 | Rust | 2026-06-30 | Self-hosted, on-premises AI coding assistant and open-source alternative to GitHub Copilot |
| 21 | Pythagora-io/gpt-pilot | 33.7k | -7 | Python | 2026-06-18 | AI developer that builds apps step by step from a spec, the engine behind the Pythagora extension; unmaintained upstream |
| 22 | Twigpine/openclaude | 33.6k | +95 | TypeScript | 2026-09-29 | Terminal coding-agent CLI that works with cloud and local model providers, with MCP support and a VS Code extension |
| 23 | charmbracelet/crush | 28.5k | +154 | Go | 2026-10-02 | Terminal coding agent that switches LLMs mid-session and uses LSPs and MCP servers for context |
| 24 | QwenLM/qwen-code | 28.3k | +144 | TypeScript | 2026-10-03 | Open-source terminal coding agent with IDE and desktop integrations that works with hosted or local models |
| 25 | Kilo-Org/kilocode | 27.5k | +56 | TypeScript | 2026-10-03 | Open-source coding agent for VS Code, JetBrains, and the CLI that works with hundreds of models |
| 26 | xai-org/grok-build | 27.2k | +93 | Rust | 2026-09-29 | xAI's terminal coding agent, a full-screen Rust TUI that also runs headless or inside editors over ACP |
| 27 | onlook-dev/onlook | 26.8k | +36 | TypeScript | 2026-08-25 | Open-source visual design tool for editing React apps with AI directly on top of your codebase |
| 28 | pingdotgg/t3code | 24.4k | — | TypeScript | 2026-10-03 | Control surface for coding agents already on your machine, such as Claude Code, Codex, and OpenCode, from desktop, web, or mobile |
| 29 | dyad-sh/dyad | 21.6k | +38 | TypeScript | 2026-10-01 | Local, open-source desktop AI app builder that uses your own API keys |
| 30 | PrimeIntellect-ai/prime-agent | 21.5k | +198 | Rust | 2026-10-03 | Terminal coding and research agent driven from a persistent Python REPL, with subagents and a self-refining harness |
| 31 | SWE-agent/SWE-agent | 20.5k | +63 | Python | 2026-09-28 | Research agent that lets a language model autonomously fix GitHub issues; superseded upstream by mini-swe-agent |
| 32 | 1jehuang/jcode | 20.3k | +146 | Rust | 2026-10-03 | Terminal coding agent harness with a TUI and SDK, built for low memory use |
| 33 | getpaseo/paseo | 19.3k | +715 | TypeScript | 2026-10-03 | Self-hosted daemon for orchestrating Claude Code, Codex, OpenCode, and other agents, driven from desktop, mobile, web, or CLI |
| 34 | avante-corp/avante.nvim | 18.2k | +1 | Lua | 2026-10-02 | Neovim plugin that brings Cursor-style AI editing to Neovim, applying suggestions directly to source files |
| 35 | HKUDS/DeepCode | 16.7k | +18 | Python | 2026-09-28 | Multi-agent coding tool with terminal, desktop, and web clients sharing one local service |
| 36 | XiaomiMiMo/MiMo-Code | 13.6k | — | TypeScript | 2026-09-30 | Terminal coding assistant from Xiaomi with persistent project memory, subagents, and any OpenAI-compatible provider |
| 37 | The-PR-Agent/pr-agent | 13.2k | — | Python | 2026-10-02 | Pull request review agent for GitHub, GitLab, Bitbucket, and more, run as a CLI, GitHub Action, or webhook |
| 38 | CodebuffAI/freebuff | 13.1k | +361 | TypeScript | 2026-10-03 | Free, ad-supported coding agent available as a CLI, desktop app, web app, and cloud agent for GitHub repos |
| 39 | MoonshotAI/kimi-cli | 11.4k | +4 | Python | 2026-09-22 | Moonshot's legacy Python terminal agent for Kimi models; archived upstream and succeeded by Kimi Code CLI |
| 40 | github/copilot-cli | 11.2k | +25 | Shell | 2026-10-02 | GitHub Copilot's terminal coding agent, with built-in GitHub integration, MCP support, and per-action approval |
| 41 | langchain-ai/open-swe | 10.8k | — | Python | 2026-10-03 | Asynchronous coding agent that investigates issues, opens pull requests, and reviews them, built on LangChain Deep Agents |
| 42 | Kuberwastaken/claurst | 10.3k | +3 | Rust | 2026-09-02 | Multi-provider terminal coding agent in Rust, begun as a clean-room reimplementation of Claude Code; beta |
How it works
data/agents.ymlis the curated source list.- A daily GitHub Actions workflow (
.github/workflows/update.yml) runs the Go updater. - The updater fetches live repo metadata via the GitHub GraphQL API in one batched query.
- Star counts are appended to
data/history.jsonlfor 7-day delta computation. - This
README.mdanddata/metadata.jsonare regenerated and committed back to the repo. - That push triggers Cloudflare Pages, which runs
go run . -buildto render the dashboard from the committed data — no API token needed at build time. See docs/DEPLOY.md.
Δ7d: Change in stars over the past 7 days; — means fewer than 7 days of history.
Contributing
Inclusion criteria — a repo belongs on this list when it is:
- A developer tool built around AI — something a developer uses to build software, where an LLM is central to what it does. That covers agents that write, edit, or review code; agent development environments that run and coordinate them; and AI-assisted editors, terminals, and review tools. Libraries, SDKs, model weights, and prompt or skill collections are out of scope — this list ranks tools you use, not building blocks you import. General-purpose assistants and multi-agent "digital workforce" apps are out as well, even when one of their agents can write code — the tool itself has to be aimed at building software.
- At least 1,000 GitHub stars — a hard floor, enforced by the updater. An entry below it is dropped from the ranking automatically, so the threshold is never waived for an individual entry.
- Open source and actively maintained — no push in 6 months means the entry is dropped. The daily run flags anything past 3 months for review, so the ranking reflects tools people can actually use today.
Archived or abandoned repos are kept only when they are historically significant, and are marked as such in data/agents.yml.
Add an agent to data/agents.yml:
agents:
- owner: github-username-or-org
repo: repository-name
description: Terminal coding agent that edits files and runs commands in your repo
tags: [terminal, byo-model, interactive, community]
Tags describe a tool across five facets — the dashboard filters on them, OR within a facet and AND across facets:
| Facet | Tags |
|---|---|
| Surface (at least one) | terminal · editor-plugin · ide · desktop · web · self-hosted |
| Model access | byo-model · single-vendor · local-models |
| Workflow | interactive · autonomous · review · app-builder · research · orchestration |
| Integration | mcp · acp · headless |
| Origin (at most one) | vendor · community |
Apply a tag only when the repo's own README, docs, or topics back it up. See docs/CONTRIBUTING.md for what each tag means.
Open a PR. The next daily run picks it up automatically.
PRs are validated automatically by CI (go run . -check): owner/repo must be non-empty and look like a real GitHub slug, every tag must come from the vocabulary above with at least one surface tag and at most one origin tag, an optional description must be one line of at most 140 characters without |, and duplicates (case-insensitive) are rejected.
License
Apache-2.0