login
RepoCritics — Review. Share. Archive. Every open-source repo.

thedotmack/claude-mem

Wiki: thedotmack/claude-mem

Source: https://github.com/thedotmack/claude-mem

Last synced 2026-06-02 · 226 words · Edit wiki on GitHub →

thedotmack/claude-mem

A memory layer for Claude Code — persistent context across sessions. Brand-attached to "claude" without explicit Anthropic endorsement; verify before integration.

What it is

A TypeScript tool that adds session-spanning memory to Claude Code. Stores context locally, surfaces it on next session start. Operates in the broader Claude-Code-ecosystem niche of "make my agent remember across runs". The repo name's "claude-" prefix is a brand association, not an endorsement signal — third-party projects that brand themselves around Anthropic products are common and not automatically vetted.

Key features

  • Persistent memory layer for Claude Code sessions.
  • TypeScript implementation.
  • MIT-licensed.

Tech stack

  • TypeScript primary.

When to reach for it

  • You're a Claude Code user wanting persistent memory.
  • You're evaluating third-party memory-augmentation options.

When not to reach for it

  • You want vendor-supported memory — Claude Code's built-in memory + project files (CLAUDE.md, AGENTS.md) are the first-party path.
  • You're risk-averse about installing third-party Claude-branded tools without verifying the actual implementation.

Maturity signal

Active in the agent-skills wave. The "claude-" prefix in the repo name is a common pattern; not an endorsement signal — verify against Anthropic's official channels before treating brand attachment as authorship.

Alternatives

  • Claude Code's built-in project memory + CLAUDE.md.
  • anthropics/skills first-party.

Tags

artificial-intelligence, agent, claude, claude-code, typescript, memory, mit-license