\n\n
![\"cursor-memory\"](\"https://github.com/user-attachments/assets/1ebad884-73d7-423e-b04d-86b74fdd44e4\")
\n\n# cursor-memory\n\n**Persistent, searchable memory for Cursor AI — You control what your AI remembers.**\n\n[](https://opensource.org/licenses/MIT)\n[](https://nodejs.org)\n[](https://modelcontextprotocol.io)\n[](https://www.npmjs.com/package/cursor-memory)\n[](https://www.npmjs.com/package/cursor-memory)\n\n[The Problem](#-the-problem) • [Quick Demo](#-see-it-in-action) • [Installation](#-installation) • [Commands](#-commands) • [How It Works](#-how-it-works) • [Troubleshooting](#-troubleshooting)\n\n
\n \n | 🎛️ You control what's saved | \n Nothing gets saved without you triggering /memo. No background processes, no noise. | \n
\n \n | ⚡ Context-aware auto-search | \n AI detects when your question refers to past context and automatically searches your memories — no command needed. /recall is available as a manual fallback. | \n
\n \n | 🔍 Finds what you mean, not what you type | \n Hybrid FTS5 keyword + vector semantic search. Searches by meaning, not just exact words. | \n
\n \n | 🌍 Cross-language | \n Save in any language, search in any language. Multilingual E5 — 100+ languages, fully cross-lingual. | \n
\n \n | 📝 Structured summaries | \n AI generates organized memos with Decisions → Key Details → Context → Next Steps — not raw text dumps. | \n
\n \n | 📄 Handles long content | \n Long discussions are automatically split into overlapping chunks — every section is searchable, nothing gets lost. | \n
\n \n | 📁 Global + per-repo scope | \n Global memories visible everywhere. Repo memories isolated per project — one repo never sees another's context. | \n
\n \n | 🧠 Choose your model | \n Small (~50MB), Medium (~115MB), or Large (~270MB) — pick the size that fits your machine. | \n
\n \n | 🔒 Fully private, runs offline | \n No cloud. No API keys. No telemetry. Everything stays on your machine. | \n
\n
\n\n---\n\n## 📦 Installation\n\n### Prerequisites\n\n- **Node.js** — Node 20, 22, or 24 LTS recommended ([download](https://nodejs.org))\n - Native modules ship prebuilt binaries for these versions, no build tools required.\n- **C++ compiler** *(optional)* — only needed as a fallback for non-LTS Node versions (Xcode CLI on Mac, `build-essential` on Ubuntu, VS Build Tools on Windows).\n\n### ⚡ 2-minute setup\n\n```bash\n# 1. Install globally\nnpm install -g cursor-memory\n\n# 2. Setup — downloads model, configures Cursor automatically\ncursor-memory setup\n\n# 3. Restart Cursor — done 🎉\n```\n\nThe CLI handles everything:\n- 📥 Downloads the embedding model\n- ⚙️ Configures MCP server for Cursor\n- 📋 Sets up AI behavior rules\n\n### 🤖 Choose your model\n\n| Model | Size | RAM | Best for |\n|-------|------|-----|----------|\n| Small | ~50MB | ~200MB | Lightweight, fast |\n| Medium | ~115MB | ~500MB | Good balance |\n| **Large** ⭐ | ~270MB | ~1GB | Best accuracy (recommended) |\n\nAll models support **100+ languages** and run **fully offline** after download.\n\n---\n\n## 💬 Commands\n\n### In Cursor Chat\n\nThree commands. That's it.\n\n| Command | What it does |\n|---------|-------------|\n| `/memo` or `/memo [text]` | 💾 With text → saves directly. Without → AI summarizes the conversation into a structured memo |\n| `/recall [query]` | 🔍 Searches your memories by keyword + semantic meaning |\n| `/forget [query]` | 🗑️ Searches → previews matches → confirms before deleting |\n\n> AI detects when your question refers to past context and automatically searches your memories — no command needed. `/recall` is available as a manual fallback.\n\n\n\n**Built with ❤️ for developers who are tired of repeating themselves to AI**\n\n[](https://www.star-history.com/?repos=tranhuucanh%2Fcursor-memory&type=date&legend=top-left)\n\n[⬆ Back to top](#-cursor-memory)\n\n
\n","readmeFilename":"README.md"}