![\"AgentKits](\"https://raw.githubusercontent.com/aitytech/agentkits-memory/main/assets/logo.svg\"){kind=logo}
\n\n \n
\n by AityTech\n
\n\n\n ![\"npm\"](\"https://img.shields.io/npm/v/@aitytech/agentkits-memory.svg\")
\n ![\"License\"](\"https://img.shields.io/badge/license-MIT-blue.svg\"){kind=icon}
\n ![\"Claude](\"https://img.shields.io/badge/Claude_Code-Compatible-blueviolet\"){kind=icon}
![\"Cursor\"](\"https://img.shields.io/badge/Cursor-Compatible-blue\"){kind=icon}
![\"Windsurf\"](\"https://img.shields.io/badge/Windsurf-Compatible-cyan\"){kind=icon}
![\"Cline\"](\"https://img.shields.io/badge/Cline-Compatible-orange\"){kind=icon}
![\"OpenCode\"](\"https://img.shields.io/badge/OpenCode-Compatible-green\"){kind=icon}
\n ![\"Tests\"](\"https://img.shields.io/badge/tests-970_passed-brightgreen\"){kind=icon}
![\"Coverage\"](\"https://img.shields.io/badge/coverage-91%25-brightgreen\"){kind=icon}
\n Persistent Memory System for AI Coding Assistants\n
\n\n\n Your AI assistant forgets everything between sessions. AgentKits Memory fixes that.
\n Decisions, patterns, errors, and context — all persisted locally via MCP.\n
\n Website •\n Docs •\n Quick Start •\n How It Works •\n Platforms •\n CLI •\n Web Viewer\n
\n\n\n English · 简体中文 · 日本語 · 한국어 · Español · Deutsch · Français · Português · Tiếng Việt · Русский · العربية\n
\n\n---\n\n## Features\n\n| Feature | Benefit |\n|---------|---------|\n| **100% Local** | All data stays on your machine. No cloud, no API keys, no accounts |\n| **Blazing Fast** | Native SQLite (better-sqlite3) = instant queries, zero latency |\n| **Zero Config** | Works out of the box. No database setup required |\n| **Multi-Platform** | Claude Code, Cursor, Windsurf, Cline, OpenCode — one setup command |\n| **MCP Server** | 9 tools: save, search, timeline, details, recall, list, update, delete, status |\n| **Auto-Capture** | Hooks capture session context, tool usage, summaries automatically |\n| **AI Enrichment** | Background workers enrich observations with AI-generated summaries |\n| **Vector Search** | sqlite-vec semantic similarity with multilingual embeddings (100+ languages) |\n| **Web Viewer** | Browser UI to view, search, add, edit, delete memories |\n| **3-Layer Search** | Progressive disclosure saves ~87% tokens vs fetching everything |\n| **Lifecycle Mgmt** | Auto-compress, archive, and clean up old sessions |\n| **Export/Import** | Backup and restore memories as JSON |\n\n---\n\n## How It Works\n\n```\nSession 1: \"Use JWT for auth\" Session 2: \"Add login endpoint\"\n┌──────────────────────────┐ ┌──────────────────────────┐\n│ You code with AI... │ │ AI already knows: │\n│ AI makes decisions │ │ ✓ JWT auth decision │\n│ AI encounters errors │ ───► │ ✓ Error solutions │\n│ AI learns patterns │ saved │ ✓ Code patterns │\n│ │ │ ✓ Session context │\n└──────────────────────────┘ └──────────────────────────┘\n │ ▲\n ▼ │\n .claude/memory/memory.db ──────────────────┘\n (SQLite, 100% local)\n```\n\n1. **Setup once** — `npx @aitytech/agentkits-memory` configures your platform\n2. **Auto-capture** — Hooks record decisions, tool usage, and summaries as you work\n3. **Context injection** — Next session starts with relevant history from past sessions\n4. **Background processing** — Workers enrich observations with AI, generate embeddings, compress old data\n5. **Search anytime** — AI uses MCP tools (`memory_search` → `memory_details`) to find past context\n\nAll data stays in `.claude/memory/memory.db` on your machine. No cloud. No API keys required.\n\n---\n\n## Design Decisions That Matter\n\nMost memory tools scatter data across markdown files, require Python runtimes, or send your code to external APIs. AgentKits Memory makes fundamentally different choices:\n\n| Design Choice | Why It Matters |\n|---------------|----------------|\n| **Single SQLite database** | One file (`memory.db`) holds everything — memories, sessions, observations, embeddings. No scattered files to sync, no merge conflicts, no orphaned data. Backup = copy one file |\n| **Native Node.js, zero Python** | Runs wherever Node runs. No conda, no pip, no virtualenv. Same language as your MCP server — one `npx` command, done |\n| **Token-efficient 3-layer search** | Search index first (~50 tokens/result), then timeline context, then full details. Only fetch what you need. Other tools dump entire memory files into context, burning tokens on irrelevant content |\n| **Auto-capture via hooks** | Decisions, patterns, and errors are recorded as they happen — not after you remember to save them. Session context injection happens automatically on next session start |\n| **Local embeddings, no API calls** | Vector search uses a local ONNX model (multilingual-e5-small). Semantic search works offline, costs nothing, and supports 100+ languages |\n| **Background workers** | AI enrichment, embedding generation, and compression run asynchronously. Your coding flow is never blocked |\n| **Multi-platform from day one** | One `--platform=all` flag configures Claude Code, Cursor, Windsurf, Cline, and OpenCode simultaneously. Same memory database, different editors |\n| **Structured observation data** | Tool usage is captured with type classification (read/write/execute/search), file tracking, intent detection, and AI-generated narratives — not raw text dumps |\n| **No process leaks** | Background workers self-terminate after 5 minutes, use PID-based lock files with stale-lock cleanup, and handle SIGTERM/SIGINT gracefully. No zombie processes, no orphaned workers |\n| **No memory leaks** | Hooks run as short-lived processes (not long-running daemons). Database connections close on shutdown. Embedding subprocess has bounded respawn (max 2), pending request timeouts, and graceful cleanup of all timers and queues |\n\n---\n\n## Web Viewer\n\nView and manage your memories through a modern web interface.\n\n```bash\nnpx @aitytech/agentkits-memory web\n```\n\nThen open **http://localhost:1905** in your browser.\n\n### Session List\n\nBrowse all sessions with timeline view and activity details.\n\n\n\n### Memory List\n\nBrowse all stored memories with search and namespace filtering.\n\n\n\n### Add Memory\n\nCreate new memories with key, namespace, type, content, and tags.\n\n\n\n### Memory Details\n\nView full memory details with edit and delete options.\n\n\n\n### Manage Embeddings\n\nGenerate and manage vector embeddings for semantic search.\n\n\n\n---\n\n## Quick Start\n\n### Option 1: Claude Code Plugin Marketplace (Recommended for Claude Code)\n\nInstall as a plugin with one command — no manual configuration needed:\n\n```bash\n/plugin marketplace add aitytech/agentkits-memory\n/plugin install agentkits-memory@agentkits-memory\n```\n\nThis installs hooks, MCP server, and memory workflow skill automatically. Restart Claude Code after installation.\n\n### Option 2: Automated Setup (All Platforms)\n\n```bash\nnpx @aitytech/agentkits-memory\n```\n\nThis auto-detects your platform and configures everything: MCP server, hooks (Claude Code/OpenCode), rules files (Cursor/Windsurf/Cline), and downloads the embedding model.\n\n**Target a specific platform:**\n\n```bash\nnpx @aitytech/agentkits-memory --platform=cursor\nnpx @aitytech/agentkits-memory --platform=windsurf,cline\nnpx @aitytech/agentkits-memory --platform=all\n```\n\n### Option 3: Manual MCP Configuration\n\nIf you prefer manual setup, add to your MCP config:\n\n```json\n{\n \"mcpServers\": {\n \"memory\": {\n \"command\": \"npx\",\n \"args\": [\"-y\", \"@aitytech/agentkits-memory\", \"server\"]\n }\n }\n}\n```\n\nConfig file locations:\n- **Claude Code**: `.claude/settings.json` (embedded in `mcpServers` key)\n- **Cursor**: `.cursor/mcp.json`\n- **Windsurf**: `.windsurf/mcp.json`\n- **Cline / OpenCode**: `.mcp.json` (project root)\n\n### 3. MCP Tools\n\nOnce configured, your AI assistant can use these tools:\n\n| Tool | Description |\n|------|-------------|\n| `memory_status` | Check memory system status (call first!) |\n| `memory_save` | Save decisions, patterns, errors, or context |\n| `memory_search` | **[Step 1]** Search index — lightweight IDs + titles (~50 tokens/result) |\n| `memory_timeline` | **[Step 2]** Get temporal context around a memory |\n| `memory_details` | **[Step 3]** Get full content for specific IDs |\n| `memory_recall` | Quick topic overview — grouped summary |\n| `memory_list` | List recent memories |\n| `memory_update` | Update existing memory content or tags |\n| `memory_delete` | Remove outdated memories |\n\n---\n\n## Progressive Disclosure (Token-Efficient Search)\n\nAgentKits Memory uses a **3-layer search pattern** that saves ~70% tokens compared to fetching full content upfront.\n\n### How It Works\n\n```\n┌─────────────────────────────────────────────────────────────┐\n│ Step 1: memory_search │\n│ Returns: IDs, titles, tags, scores (~50 tokens/item) │\n│ → Review index, pick relevant memories │\n└─────────────────────────────────────────────────────────────┘\n ↓\n┌─────────────────────────────────────────────────────────────┐\n│ Step 2: memory_timeline (optional) │\n│ Returns: Context ±30 minutes around memory │\n│ → Understand what happened before/after │\n└─────────────────────────────────────────────────────────────┘\n ↓\n┌─────────────────────────────────────────────────────────────┐\n│ Step 3: memory_details │\n│ Returns: Full content for selected IDs only │\n│ → Fetch only what you actually need │\n└─────────────────────────────────────────────────────────────┘\n```\n\n### Example Workflow\n\n```typescript\n// Step 1: Search - get lightweight index\nmemory_search({ query: \"authentication\" })\n// → Returns: [{ id: \"abc\", title: \"JWT pattern...\", score: 85% }]\n\n// Step 2: (Optional) See temporal context\nmemory_timeline({ anchor: \"abc\" })\n// → Returns: What happened before/after this memory\n\n// Step 3: Get full content only for what you need\nmemory_details({ ids: [\"abc\"] })\n// → Returns: Full content for selected memory\n```\n\n### Token Savings\n\n| Approach | Tokens Used |\n|----------|-------------|\n| **Old:** Fetch all content | ~500 tokens × 10 results = 5000 tokens |\n| **New:** Progressive disclosure | 50 × 10 + 500 × 2 = 1500 tokens |\n| **Savings** | **70% reduction** |\n\n---\n\n## CLI Commands\n\n```bash\n# One-command setup (auto-detects platform)\nnpx @aitytech/agentkits-memory\nnpx @aitytech/agentkits-memory setup --platform=cursor # specific platform\nnpx @aitytech/agentkits-memory setup --platform=all # all platforms\nnpx @aitytech/agentkits-memory setup --force # re-install/update\n\n# Start MCP server\nnpx @aitytech/agentkits-memory server\n\n# Web viewer (port 1905)\nnpx @aitytech/agentkits-memory web\n\n# Terminal viewer\nnpx @aitytech/agentkits-memory viewer\nnpx @aitytech/agentkits-memory viewer --stats\nnpx @aitytech/agentkits-memory viewer --json\n\n# Save from CLI\nnpx @aitytech/agentkits-memory save \"Use JWT with refresh tokens\" --category pattern --tags auth,security\n\n# Settings\nnpx @aitytech/agentkits-memory hook settings .\nnpx @aitytech/agentkits-memory hook settings . --reset\nnpx @aitytech/agentkits-memory hook settings . aiProvider.provider=openai aiProvider.apiKey=sk-...\n\n# Export / Import\nnpx @aitytech/agentkits-memory hook export . my-project ./backup.json\nnpx @aitytech/agentkits-memory hook import . ./backup.json\n\n# Lifecycle management\nnpx @aitytech/agentkits-memory hook lifecycle . --compress-days=7 --archive-days=30\nnpx @aitytech/agentkits-memory hook lifecycle-stats .\n```\n\n---\n\n## Programmatic Usage\n\n```typescript\nimport { ProjectMemoryService } from '@aitytech/agentkits-memory';\n\nconst memory = new ProjectMemoryService({\n baseDir: '.claude/memory',\n dbFilename: 'memory.db',\n});\nawait memory.initialize();\n\n// Store a memory\nawait memory.storeEntry({\n key: 'auth-pattern',\n content: 'Use JWT with refresh tokens for authentication',\n namespace: 'patterns',\n tags: ['auth', 'security'],\n});\n\n// Query memories\nconst results = await memory.query({\n type: 'hybrid',\n namespace: 'patterns',\n content: 'authentication',\n limit: 10,\n});\n\n// Get by key\nconst entry = await memory.getByKey('patterns', 'auth-pattern');\n```\n\n---\n\n## Auto-Capture Hooks\n\nHooks automatically capture your AI coding sessions (Claude Code and OpenCode only):\n\n| Hook | Trigger | Action |\n|------|---------|--------|\n| `context` | Session Start | Injects previous session context + memory status |\n| `session-init` | User Prompt | Initializes/resumes session, records prompts |\n| `observation` | After Tool Use | Captures tool usage with intent detection |\n| `summarize` | Session End | Generates structured session summary |\n| `user-message` | Session Start | Displays memory status to user (stderr) |\n\nSetup hooks:\n```bash\nnpx @aitytech/agentkits-memory\n```\n\n**What gets captured automatically:**\n- File reads/writes with paths\n- Code changes as structured diffs (before → after)\n- Developer intent (bugfix, feature, refactor, investigation, etc.)\n- Session summaries with decisions, errors, and next steps\n- Multi-prompt tracking within sessions\n\n---\n\n## Multi-Platform Support\n\n| Platform | MCP | Hooks | Rules File | Setup |\n|----------|-----|-------|------------|-------|\n| **Claude Code** | `.claude/settings.json` | ✅ Full | CLAUDE.md (skill) | `--platform=claude-code` |\n| **Cursor** | `.cursor/mcp.json` | — | `.cursorrules` | `--platform=cursor` |\n| **Windsurf** | `.windsurf/mcp.json` | — | `.windsurfrules` | `--platform=windsurf` |\n| **Cline** | `.mcp.json` | — | `.clinerules` | `--platform=cline` |\n| **OpenCode** | `.mcp.json` | ✅ Full | — | `--platform=opencode` |\n\n- **MCP Server** works with all platforms (memory tools via MCP protocol)\n- **Hooks** provide auto-capture on Claude Code and OpenCode\n- **Rules files** teach Cursor/Windsurf/Cline the memory workflow\n- **Memory data** always stored in `.claude/memory/` (single source of truth)\n\n---\n\n## Background Workers\n\nAfter each session, background workers process queued tasks:\n\n| Worker | Task | Description |\n|--------|------|-------------|\n| `embed-session` | Embeddings | Generate vector embeddings for semantic search |\n| `enrich-session` | AI Enrichment | Enrich observations with AI-generated summaries, facts, concepts |\n| `compress-session` | Compression | Compress old observations (10:1–25:1) and generate session digests (20:1–100:1) |\n\nWorkers run automatically after session end. Each worker:\n- Processes up to 200 items per run\n- Uses lock files to prevent concurrent execution\n- Auto-terminates after 5 minutes (prevents zombies)\n- Retries failed tasks up to 3 times\n\n---\n\n## AI Provider Configuration\n\nAI enrichment uses pluggable providers. Default is `claude-cli` (no API key needed).\n\n| Provider | Type | Default Model | Notes |\n|----------|------|---------------|-------|\n| **Claude CLI** | `claude-cli` | `haiku` | Uses `claude --print`, no API key needed |\n| **OpenAI** | `openai` | `gpt-4o-mini` | Any OpenAI model |\n| **Google Gemini** | `gemini` | `gemini-2.0-flash` | Google AI Studio key |\n| **OpenRouter** | `openai` | any | Set `baseUrl` to `https://openrouter.ai/api/v1` |\n| **GLM (Zhipu)** | `openai` | any | Set `baseUrl` to `https://open.bigmodel.cn/api/paas/v4` |\n| **Ollama** | `openai` | any | Set `baseUrl` to `http://localhost:11434/v1` |\n\n### Option 1: Environment Variables\n\n```bash\n# OpenAI\nexport AGENTKITS_AI_PROVIDER=openai\nexport AGENTKITS_AI_API_KEY=sk-...\n\n# Google Gemini\nexport AGENTKITS_AI_PROVIDER=gemini\nexport AGENTKITS_AI_API_KEY=AIza...\n\n# OpenRouter (uses OpenAI-compatible format)\nexport AGENTKITS_AI_PROVIDER=openai\nexport AGENTKITS_AI_API_KEY=sk-or-...\nexport AGENTKITS_AI_BASE_URL=https://openrouter.ai/api/v1\nexport AGENTKITS_AI_MODEL=anthropic/claude-3.5-haiku\n\n# Local Ollama (no API key needed)\nexport AGENTKITS_AI_PROVIDER=openai\nexport AGENTKITS_AI_BASE_URL=http://localhost:11434/v1\nexport AGENTKITS_AI_MODEL=llama3.2\n\n# Disable AI enrichment entirely\nexport AGENTKITS_AI_ENRICHMENT=false\n```\n\n### Option 2: Persistent Settings\n\n```bash\n# Saved to .claude/memory/settings.json — persists across sessions\nnpx @aitytech/agentkits-memory hook settings . aiProvider.provider=openai aiProvider.apiKey=sk-...\nnpx @aitytech/agentkits-memory hook settings . aiProvider.provider=gemini aiProvider.apiKey=AIza...\nnpx @aitytech/agentkits-memory hook settings . aiProvider.baseUrl=https://openrouter.ai/api/v1\n\n# View current settings\nnpx @aitytech/agentkits-memory hook settings .\n\n# Reset to defaults\nnpx @aitytech/agentkits-memory hook settings . --reset\n```\n\n> **Priority:** Environment variables override settings.json. Settings.json overrides defaults.\n\n---\n\n## Lifecycle Management\n\nManage memory growth over time:\n\n```bash\n# Compress observations older than 7 days, archive sessions older than 30 days\nnpx @aitytech/agentkits-memory hook lifecycle . --compress-days=7 --archive-days=30\n\n# Also auto-delete archived sessions older than 90 days\nnpx @aitytech/agentkits-memory hook lifecycle . --compress-days=7 --archive-days=30 --delete --delete-days=90\n\n# View lifecycle statistics\nnpx @aitytech/agentkits-memory hook lifecycle-stats .\n```\n\n| Stage | What Happens |\n|-------|-------------|\n| **Compress** | AI-compresses observations, generates session digests |\n| **Archive** | Marks old sessions as archived (excluded from context) |\n| **Delete** | Removes archived sessions (opt-in, requires `--delete`) |\n\n---\n\n## Export / Import\n\nBackup and restore your project memories:\n\n```bash\n# Export all sessions for a project\nnpx @aitytech/agentkits-memory hook export . my-project ./backup.json\n\n# Import from backup (deduplicates automatically)\nnpx @aitytech/agentkits-memory hook import . ./backup.json\n```\n\nExport format includes sessions, observations, prompts, and summaries.\n\n---\n\n## Memory Categories\n\n| Category | Use Case |\n|----------|----------|\n| `decision` | Architecture decisions, tech stack picks, trade-offs |\n| `pattern` | Coding conventions, project patterns, recurring approaches |\n| `error` | Bug fixes, error solutions, debugging insights |\n| `context` | Project background, team conventions, environment setup |\n| `observation` | Auto-captured session observations |\n\n---\n\n## Storage\n\nMemories are stored in `.claude/memory/` within your project directory.\n\n```\n.claude/memory/\n├── memory.db # SQLite database (all data)\n├── memory.db-wal # Write-ahead log (temp)\n├── settings.json # Persistent settings (AI provider, context config)\n└── embeddings-cache/ # Cached vector embeddings\n```\n\n---\n\n## CJK Language Support\n\nAgentKits Memory has **automatic CJK support** for Chinese, Japanese, and Korean text search.\n\n### Zero Configuration\n\nWhen `better-sqlite3` is installed (default), CJK search works automatically:\n\n```typescript\nimport { ProjectMemoryService } from '@aitytech/agentkits-memory';\n\nconst memory = new ProjectMemoryService('.claude/memory');\nawait memory.initialize();\n\n// Store CJK content\nawait memory.storeEntry({\n key: 'auth-pattern',\n content: '認証機能の実装パターン - JWT with refresh tokens',\n namespace: 'patterns',\n});\n\n// Search in Japanese, Chinese, or Korean - it just works!\nconst results = await memory.query({\n type: 'hybrid',\n content: '認証機能',\n});\n```\n\n### How It Works\n\n- **Native SQLite**: Uses `better-sqlite3` for maximum performance\n- **Trigram tokenizer**: FTS5 with trigram creates 3-character sequences for CJK matching\n- **Smart fallback**: Short CJK queries (< 3 chars) automatically use LIKE search\n- **BM25 ranking**: Relevance scoring for search results\n\n### Advanced: Japanese Word Segmentation\n\nFor advanced Japanese with proper word segmentation, optionally use lindera:\n\n```typescript\nimport { createJapaneseOptimizedBackend } from '@aitytech/agentkits-memory';\n\nconst backend = createJapaneseOptimizedBackend({\n databasePath: '.claude/memory/memory.db',\n linderaPath: './path/to/liblindera_sqlite.dylib',\n});\n```\n\nRequires [lindera-sqlite](https://github.com/lindera/lindera-sqlite) build.\n\n---\n\n## API Reference\n\n### ProjectMemoryService\n\n```typescript\ninterface ProjectMemoryConfig {\n baseDir: string; // Default: '.claude/memory'\n dbFilename: string; // Default: 'memory.db'\n enableVectorIndex: boolean; // Default: false\n dimensions: number; // Default: 384\n embeddingGenerator?: EmbeddingGenerator;\n cacheEnabled: boolean; // Default: true\n cacheSize: number; // Default: 1000\n cacheTtl: number; // Default: 300000 (5 min)\n}\n```\n\n### Methods\n\n| Method | Description |\n|--------|-------------|\n| `initialize()` | Initialize the memory service |\n| `shutdown()` | Shutdown and persist changes |\n| `storeEntry(input)` | Store a memory entry |\n| `get(id)` | Get entry by ID |\n| `getByKey(namespace, key)` | Get entry by namespace and key |\n| `update(id, update)` | Update an entry |\n| `delete(id)` | Delete an entry |\n| `query(query)` | Query entries with filters |\n| `semanticSearch(content, k)` | Semantic similarity search |\n| `count(namespace?)` | Count entries |\n| `listNamespaces()` | List all namespaces |\n| `getStats()` | Get statistics |\n\n---\n\n## Code Quality\n\nAgentKits Memory is thoroughly tested with **970 unit tests** across 21 test suites.\n\n| Metric | Coverage |\n|--------|----------|\n| **Statements** | 90.29% |\n| **Branches** | 80.85% |\n| **Functions** | 90.54% |\n| **Lines** | 91.74% |\n\n### Test Categories\n\n| Category | Tests | What's Covered |\n|----------|-------|----------------|\n| Core Memory Service | 56 | CRUD, search, pagination, categories, tags, import/export |\n| SQLite Backend | 65 | Schema, migrations, FTS5, transactions, error handling |\n| sqlite-vec Vector Index | 47 | Insert, search, delete, persistence, edge cases |\n| Hybrid Search | 44 | FTS + vector fusion, scoring, ranking, filters |\n| Token Economics | 27 | 3-layer search budgets, truncation, optimization |\n| Embedding System | 63 | Cache, subprocess, local models, CJK support |\n| Hook System | 502 | Context, session-init, observation, summarize, AI enrichment, service lifecycle, queue workers, adapters, types |\n| MCP Server | 48 | All 9 MCP tools, validation, error responses |\n| CLI | 34 | Platform detection, rules generation |\n| Integration | 84 | End-to-end flows, embedding integration, multi-session |\n\n```bash\n# Run tests\nnpm test\n\n# Run with coverage\nnpm run test:coverage\n```\n\n---\n\n## Requirements\n\n- **Node.js LTS**: 18.x, 20.x, or 22.x (recommended)\n- MCP-compatible AI coding assistant\n\n### Node.js Version Notes\n\nThis package uses `better-sqlite3` which requires native binaries. **Prebuilt binaries are available for LTS versions only**.\n\n| Node Version | Status | Notes |\n|--------------|--------|-------|\n| 18.x LTS | ✅ Works | Prebuilt binaries |\n| 20.x LTS | ✅ Works | Prebuilt binaries |\n| 22.x LTS | ✅ Works | Prebuilt binaries |\n| 19.x, 21.x, 23.x | ⚠️ Requires build tools | No prebuilt binaries |\n\n### Using Non-LTS Versions (Windows)\n\nIf you must use a non-LTS version (19, 21, 23), install build tools first:\n\n**Option 1: Visual Studio Build Tools**\n```powershell\n# Download and install from:\n# https://visualstudio.microsoft.com/visual-cpp-build-tools/\n# Select \"Desktop development with C++\" workload\n```\n\n**Option 2: windows-build-tools (npm)**\n```powershell\nnpm install --global windows-build-tools\n```\n\n**Option 3: Chocolatey**\n```powershell\nchoco install visualstudio2022-workload-vctools\n```\n\nSee [node-gyp Windows guide](https://github.com/nodejs/node-gyp#on-windows) for more details.\n\n---\n\n## AgentKits Ecosystem\n\n**AgentKits Memory** is part of the AgentKits ecosystem by AityTech - tools that make AI coding assistants smarter.\n\n| Product | Description | Link |\n|---------|-------------|------|\n| **AgentKits Engineer** | 28 specialized agents, 100+ skills, enterprise patterns | [GitHub](https://github.com/aitytech/agentkits-engineer) |\n| **AgentKits Marketing** | AI-powered marketing content generation | [GitHub](https://github.com/aitytech/agentkits-marketing) |\n| **AgentKits Memory** | Persistent memory for AI assistants (this package) | [npm](https://www.npmjs.com/package/@aitytech/agentkits-memory) |\n\n\n \n ![\"agentkits.net\"](\"https://img.shields.io/badge/Visit-agentkits.net-blue?style=for-the-badge\"){kind=icon}
\n Give your AI assistant memory that persists.\n
\n\n\n AgentKits Memory by AityTech\n
\n\n\n Star this repo if it helps your AI remember.\n
\n","readmeFilename":"README.md"}