Quick Start • How It Works • Search Tools • Documentation • Configuration • Troubleshooting • License

Claude-Mem seamlessly preserves context across sessions by automatically capturing tool usage observations, generating semantic summaries, and making them available to future sessions. This enables Claude to maintain continuity of knowledge about projects even after sessions end or reconnect.

We'd love your feedback! Test the new context injection system in feature/context-with-observations

Progressive disclosure is a layered approach to memory retrieval that mirrors how humans remember information:

Layer 1: Index (Frontloaded at session start)

• WHAT exists: Observation titles and session summaries
• COST of retrieval: Token counts for each observation
• TYPE indicators: Critical (🔴 gotcha, 🟤 decision) vs informational (🔵 how-it-works)

Layer 2: Details (Retrieved on-demand via MCP search)

• Full observation narratives when Claude needs deeper context
• Search by concept, file, type, or keyword

Layer 3: Perfect Recall (Code/transcripts)

• Source code when implementation details are needed
• Original transcripts for exact quotes

Current version (v4.2.x): Shows only session-level summaries at startup

• ✅ Minimal tokens (~800)
• ❌ Claude doesn't know what detailed observations exist
• ❌ Often re-reads code to understand past decisions

Experimental version: Shows observation index + session summaries

• ✅ Claude sees WHAT learnings exist without loading full content
• ✅ Token counts help Claude decide: "fetch details" vs "read code"
• ✅ Progressive disclosure instructions teach Claude how to use the system
• ⚠️ Higher token cost (~2,500) but potentially more efficient overall

The experimental context hook displays observations in a table format:

**src/hooks/context.ts**
| ID | Time | T | Title | Tokens |
|----|------|---|-------|--------|
| #2332 | 1:07 AM | 🔴 | Critical Bugfix: Session ID NULL Constraint | ~201 |
| #2340 | 1:10 AM | 🟠 | Remove Redundant Summary Section | ~280 |

Now Claude knows:

• A critical bugfix exists about session IDs (~201 tokens to fetch)
• A design decision exists about summary sections (~280 tokens)
• Whether to use MCP search or just read the current code

# Clone and checkout the experimental branch
git clone https://github.com/thedotmack/claude-mem.git
cd claude-mem
git checkout feature/context-with-observations

# Build the experimental version
npm install
npm run build

# Test the new context hook (from YOUR project directory)
cd /path/to/your/project
node /path/to/claude-mem/plugin/scripts/context-hook.js

# Example:
# cd ~/my-app
# node ~/claude-mem/plugin/scripts/context-hook.js

Important: Run the context hook from your project's root directory to see context specific to that project.

We want to know:

• Does Claude use MCP search more effectively?
• Do token counts influence retrieval decisions?
• Is the progressive disclosure guidance helpful or noisy?
• Does it reduce redundant code reading?

📣 Share Your Feedback: Open a GitHub Issue with your experience! Tag it with feedback: progressive-disclosure

Start a new Claude Code session in the terminal and enter the following commands:

> /plugin marketplace add thedotmack/claude-mem

> /plugin install claude-mem

Restart Claude Code. Context from previous sessions will automatically appear in new sessions.

Key Features:

• 🧠 Persistent Memory - Context survives across sessions
• 🔍 7 Search Tools - Query your project history via MCP
• 🤖 Automatic Operation - No manual intervention required
• 📊 FTS5 Search - Fast full-text search across observations
• 🔗 Citations - Reference past decisions with claude-mem:// URIs

📚 View Full Documentation - Browse markdown docs on GitHub

💻 Local Preview: Run Mintlify docs locally:

cd docs
npx mintlify dev

• Installation Guide - Quick start & advanced installation
• Usage Guide - How Claude-Mem works automatically
• MCP Search Tools - Query your project history

• Overview - System components & data flow
• Hooks - 5 lifecycle hooks explained
• Worker Service - HTTP API & PM2 management
• Database - SQLite schema & FTS5 search
• MCP Search - 7 search tools & examples

• Configuration - Environment variables & settings
• Development - Building, testing, contributing
• Troubleshooting - Common issues & solutions

┌─────────────────────────────────────────────────────────────┐
│ Session Start → Inject context from last 10 sessions       │
└─────────────────────────────────────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────────┐
│ User Prompts → Create session, save user prompts           │
└─────────────────────────────────────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────────┐
│ Tool Executions → Capture observations (Read, Write, etc.)  │
└─────────────────────────────────────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────────┐
│ Worker Processes → Extract learnings via Claude Agent SDK   │
└─────────────────────────────────────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────────┐
│ Session Ends → Generate summary, ready for next session     │
└─────────────────────────────────────────────────────────────┘

Core Components:

1. 5 Lifecycle Hooks - SessionStart, UserPromptSubmit, PostToolUse, Stop, SessionEnd
2. Worker Service - HTTP API on port 37777 managed by PM2
3. SQLite Database - Stores sessions, observations, summaries with FTS5 search
4. 7 MCP Search Tools - Query historical context with citations

See Architecture Overview for details.

Claude-Mem provides 7 specialized search tools:

1. search_observations - Full-text search across observations
2. search_sessions - Full-text search across session summaries
3. search_user_prompts - Search raw user requests
4. find_by_concept - Find by concept tags
5. find_by_file - Find by file references
6. find_by_type - Find by type (decision, bugfix, feature, etc.)
7. get_recent_context - Get recent session context

Example Queries:

search_observations with query="authentication" and type="decision"
find_by_file with filePath="worker-service.ts"
search_user_prompts with query="add dark mode"
get_recent_context with limit=5

See MCP Search Tools Guide for detailed examples.

Security:

• Fixed FTS5 injection vulnerability in search functions
• Added comprehensive test suite with 332 injection attack tests

Fixes:

• Fixed ESM/CJS compatibility for getDirname function
• Fixed Windows PowerShell compatibility in SessionStart hook
• Cross-platform dependency installation now works on Windows, macOS, and Linux

See CHANGELOG.md for complete version history.

• Node.js: 18.0.0 or higher
• Claude Code: Latest version with plugin support
• PM2: Process manager (bundled - no global install required)
• SQLite 3: For persistent storage (bundled)

• Context automatically injected when Claude starts
• No manual commands or configuration needed
• Works transparently in the background

• Search across all sessions and observations
• FTS5 full-text search for fast queries
• Citations link back to specific observations

• AI-powered extraction of learnings
• Categorized by type (decision, bugfix, feature, etc.)
• Tagged with concepts and file references

• Sessions span multiple user prompts
• Context preserved across /clear commands
• Track entire conversation threads

Model Selection:

./claude-mem-settings.sh

Environment Variables:

• CLAUDE_MEM_MODEL - AI model for processing (default: claude-sonnet-4-5)
• CLAUDE_MEM_WORKER_PORT - Worker port (default: 37777)
• CLAUDE_MEM_DATA_DIR - Data directory override (dev only)

See Configuration Guide for details.

# Clone and build
git clone https://github.com/thedotmack/claude-mem.git
cd claude-mem
npm install
npm run build

# Run tests
npm test

# Start worker
npm run worker:start

# View logs
npm run worker:logs

See Development Guide for detailed instructions.

Common Issues:

• Worker not starting → npm run worker:restart
• No context appearing → npm run test:context
• Database issues → sqlite3 ~/.claude-mem/claude-mem.db "PRAGMA integrity_check;"
• Search not working → Check FTS5 tables exist

See Troubleshooting Guide for complete solutions.

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes with tests
4. Update documentation
5. Submit a Pull Request

See Development Guide for contribution workflow.

This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).

Copyright (C) 2025 Alex Newman (@thedotmack). All rights reserved.

See the LICENSE file for full details.

What This Means:

• You can use, modify, and distribute this software freely
• If you modify and deploy on a network server, you must make your source code available
• Derivative works must also be licensed under AGPL-3.0
• There is NO WARRANTY for this software

• Documentation: docs/
• Issues: GitHub Issues
• Repository: github.com/thedotmack/claude-mem
• Author: Alex Newman (@thedotmack)

Built with Claude Agent SDK | Powered by Claude Code | Made with TypeScript