The gist

Claude-Mem is a persistent memory system for AI agents, built by Alex Newman. It solves the problem of context loss between sessions by automatically capturing tool usage, generating semantic summaries, and injecting relevant memories into new sessions. This gives AI coding assistants like Claude a continuous understanding of a project's history and state, even after being restarted.

What it does

Preserves context across development sessions by automatically capturing tool observations.
Generates semantic summaries of interactions for efficient future retrieval.
Enables natural language search of project history via a dedicated skill.
Provides a local web viewer to monitor the real-time memory stream.
Controls what context gets injected into sessions through fine-grained configuration.
Excludes sensitive information from memory storage using private tags.

How it works

Claude-Mem is installed via a CLI command and runs as a local worker service. It uses lifecycle hooks to automatically capture tool interactions during coding sessions. This data is stored in a local SQLite database, with embeddings managed by a Chroma vector database for hybrid semantic and keyword search. The tool is open-source under an AGPL-3.0 license and operates automatically in the background after initial setup.

Best for

This tool is best for developers using AI coding assistants like Claude Code or Gemini CLI who need to maintain project-specific context and memory across multiple, disconnected work sessions.

Watch out for

The project is licensed under AGPL-3.0, which requires derivative works deployed over a network to also be open-sourced. A subdirectory uses a separate, non-commercial license.