thedotmack / claude-mem
Summary
Claude-Mem is a persistent memory compression system designed for Claude Code, enabling AI agents to maintain context across sessions by automatically capturing tool usage, generating semantic summaries, and making them available for future interactions. The system, currently at version 6.5.0, operates on Solana with a contract address of 2TsmuYUrsctE57VLckZBYEEzdokUF8j8e1GavekWBAGS. Key features include persistent memory, progressive disclosure of context with token cost visibility, skill-based search, a web viewer UI at `http://localhost:37777`, and privacy controls. It supports Node.js 18.0.0+, Claude Code, Bun, uv, and SQLite 3, and is licensed under AGPL-3.0, with the `ragtime/` directory under PolyForm Noncommercial License 1.0.0.
Key takeaway
For AI Engineers developing with Claude Code, Claude-Mem offers a critical solution for maintaining project context across sessions. Implementing this system allows your AI agents to retain knowledge and improve continuity, reducing the need for repetitive context setting. Consider integrating Claude-Mem to enhance agent efficiency and enable more complex, long-running development tasks.
Key insights
Claude-Mem provides persistent, searchable context for Claude Code sessions, enhancing AI agent continuity and efficiency.
Principles
- Context survives across sessions.
- Layered memory retrieval optimizes token cost.
- Hybrid search combines semantic and keyword queries.
Method
Claude-Mem uses 5 lifecycle hooks, a worker service with an HTTP API, SQLite for storage, and a Chroma Vector Database for hybrid search, all managed by Bun.
In practice
- Install via `/plugin marketplace add thedotmack/claude-mem`.
- Use `<exclude>` tags to protect sensitive content.
- Access real-time memory stream at `http://localhost:37777`.
Topics
- Persistent Memory
- Context Management
- Claude Code
- Vector Databases
- Hybrid Search
Code references
Best for: AI Engineer, Machine Learning Engineer, Software Engineer
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by Github Trending: All languages.