thedotmack/claude-mem

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

What it solves

Claude-Mem solves the problem of "amnesia" in AI coding assistants. By default, AI tools often lose context when a chat session ends or restarts. This project ensures that your project history, past decisions, and bug fixes remain available to the AI across different sessions, allowing it to maintain continuity of knowledge.

How it works

It operates as a persistent memory layer that hooks into your AI coding environment. It automatically captures "observations"—key events like tool usage, bug fixes, and project decisions—and stores them in a local SQLite database. It uses a hybrid search approach, combining keyword and semantic search (via a vector database), to retrieve relevant past information. When you start a new session, it injects this historical context back into the AI, using a "progressive disclosure" strategy to keep token costs low by only fetching full details when necessary.

Who it’s for

It is designed for developers who use Claude Code or similar AI coding agents and want their assistants to remember previous work, project context, and technical decisions without having to manually re-explain everything in every new chat.

Highlights

  • Persistent Context: Automatically saves and retrieves project history across different sessions.
  • Token-Efficient Search: Uses a three-step workflow (search, timeline, fetch) to minimize token usage while retrieving relevant data.
  • Privacy-First: Includes support for <private> tags to exclude sensitive information from being stored.
  • Automated Integration: Easily installs via CLI and works automatically in the background without manual intervention.
  • Team Capabilities: Optional shared workspace features for team collaboration and knowledge sharing.

Related

  • Project
    supermemoryai/claude-supermemory

    claude-supermemory is a Claude Code plugin that gives the LLM coding agent persistent memory across sessions. It automatically captures conversation turns, recalls relevant past context at prompt time, and shares that memory across tools like Codex and OpenCode via a unified container tag.

  • Project
    YeonwooSung/my-claude-code-mods

    A set of Node‑based plugins (mods) for Anthropic’s Claude Code that add profiling, context analysis, timeline snapshots, loop detection, pinned notes, and model‑effort visualisation. Each mod is installable via Claude’s marketplace and includes unit tests and TypeScript type‑checking.

  • Dispatch
    Recall: Local Project Memory for Claude Code

    Recall is an offline, local-first plugin for Claude Code that captures session history and uses a TF-IDF and TextRank summarizer to create durable project memory without spending model tokens.

  • Project
    GMaN1911/claude-cognitive

    Claude Cognitive adds persistent working memory to Claude Code, using a context router that scores and injects files based on attention and a pool coordinator that shares state across multiple instances, cutting token usage and preventing duplicate work.

  • Project
    qf-studio/navigator

    A Claude Code plugin that implements context engineering to prevent session crashes and hallucinations by loading only necessary information, significantly extending session longevity.