開発ツール
cc-connect
Bridge local AI coding agents (Claude Code, Cursor, Gemini CLI, Codex) to messaging platforms (Feishu/Lark, DingTalk, Slack, Telegram, Discord, LINE, WeChat …
★ 15.7k スターUnknown
ガイドを見る →Graftは、trailhqが公開しているオープンソースの開発ツールプロジェクトです（GitHub: trailhq/Graft）。GitHubスターは9,787、フォークは885で、MITライセンスの下で公開されています。GitHubでの最終更新は2026年10月8日です。
A codebase knowledge graph that makes Claude Code, Cursor, Codex and Gemini faster, cheaper and repo-aware.
Pick the coding agent you use and the LLM provider whose key will write Graft's optional deep summaries; the structural graph itself needs no key.
Clone trailhq/Graft from GitHub (or use the install path in the run steps below) and skim its README for the parts relevant to your goal.
Install the CLI globally from npm, then run graft init in your repo to drop in the wiring for your agent.
Run graft build to generate the graph, then ask your agent a codebase question and let it follow the graph instead of grepping from scratch.
Each node explains in plain English what part of the system does and how it connects, the way a senior engineer would — not just a dump of symbols.
Every query rebuilds the graph against the working tree first (about 3ms when nothing moved), so answers include uncommitted edits with no stale index.
graft build writes a regenerable local graft/ cache (gitignored, like node_modules); you only commit the small init wiring — no database or server.
Summaries are written by any provider you choose — OpenAI, Anthropic natively, OpenRouter, Fireworks, Groq, OrcaRouter, a LiteLLM proxy or a local model — under your own key.
The core graph, graft check and graft ask are deterministic tree-sitter and never call a model, so they work offline with no key.
graft build --deep adds LLM concept nodes and per-symbol summaries; --lsp adds opt-in compiler-grade edges on top of scope-aware cross-file extraction.
Get the source from the official GitHub repository for Graft.
git clone https://github.com/trailhq/Graft.git
cd GraftInstall Graft globally from npm, exactly as the README's quick start shows (or use npx without installing).
npm install -g @nanonets/graftWire Graft into your repo for your coding agent, then build the codebase graph.
npx @nanonets/graft init
graft buildTrailhq, through the trailhq/Graft repository, with the product site at trailhq.com/graft.
Yes — the project is open source under the MIT license as recorded for this listing. You can use, modify and distribute it under the MIT terms.
The structural graph needs no API key at all. Only the optional --deep summaries call an LLM, using your own provider key, so any cost is your provider usage.
Embedding indexes need similarity search and a warm index that can go stale. Graft is a real, readable graph of linked files your agent opens and follows, rebuilt deterministically on every query — no embeddings and no server.
Clone the repository (or use the documented installer), follow the install and configure steps on this page, and run the smallest real task first — run graft build to generate the graph, then ask your agent a codebase question and let it follow the graph instead of grepping from scratch.
Bridge local AI coding agents (Claude Code, Cursor, Gemini CLI, Codex) to messaging platforms (Feishu/Lark, DingTalk, Slack, Telegram, Discord, LINE, WeChat …
Lightweight AST-based semantic code search CLI (ccc) that makes coding agents faster and cheaper
Gentle-AI configures the AI coding agents you already use: Claude Code, Cursor, OpenCode, Codex, Pi, and more. Choose persistent memory, Organic-Driven Devel…
Menu-bar tracker for your AI coding quotas — Claude, Codex, Gemini and more.
Claude Octopus orchestrates Claude, Codex, Copilot, Gemini, and Ollama on one task — and surfaces disagreement
A local-first Agent OS — drive Claude Code, Codex and OpenCode from your browser.
LibreAIカタログに掲載された開発ツールプロジェクトの中で最もスターの多いもの — Graftが合わなかった場合に検討したい、現実的な候補リストです。
The open source coding agent
The model-definition framework behind modern AI — thousands of pretrained transformers for text, vision, and audio.
Anthropic's agentic coding tool that lives in your terminal and works across your whole codebase
Cross-platform desktop all-in-one manager for Claude Code, Codex, OpenCode and other coding agents
OpenAI's lightweight coding agent that runs locally in your terminal
カタログの中でGraftとスター数が近い開発ツールプロジェクト — 前後の棚もそのまま見て回れます。
One Rust CLI for every LLM: shell assistant, chat REPL, RAG, tools and agents across 20+ providers.
An autonomous loop for Claude Code with intelligent exit detection, rate limits and a circuit breaker.
ARTEMIS turns natural-language instructions into reliable Android automation. It automates end-to-end workflows, captures logs, and integrates seamlessly wit…
Source control for agents — run coding agents side by side, track changes and query their history.
Expose your FastAPI endpoints as Model Context Protocol tools - with authentication - in a few lines of code.
✨ The agentic HTML editor — your local AI agent writes the HTML, you ship it. 🚀 75 Skills × 9 Surfaces (magazine · deck · poster · XHS / tweet · prototype · …