claude-mem

Productivity

Persistent cross-session memory: observations captured automatically, searchable forever.

by thedotmackthedotmack/claude-mem

Stars102.3k

Forks9k

Watchers322

Open issues128

LicenseApache-2.0

Last pushtoday

Install

install

$ /plugin marketplace add thedotmack/claude-mem

Update

update

$ /plugin marketplace update claude-mem

Then run /reload-plugins. Official marketplaces auto-update at startup.

When to use it

Long-running projects where context from past sessions keeps getting lost
Answering questions like: did we already solve this? how did we do X last time?
Generating project history reports and week-by-week development narratives

Commands & flags

/claude-mem:mem-search: Search past sessions in the memory database.
/claude-mem:make-plan: Create phased plans with automatic doc discovery.
/claude-mem:do: Execute a phased plan with subagents.
/claude-mem:timeline-report: Generate a narrative project history report.

How to use it

Once installed, it captures observations from every session automatically and injects relevant memory when topics recur. Query it directly:

/claude-mem:mem-search   search past sessions
/claude-mem:make-plan    phased plans with doc discovery
/claude-mem:do           execute a phased plan with subagents
/claude-mem:timeline-report   narrative project history

Why it earns its place

Claude Code forgets everything between sessions by default. claude-mem turns past work into a searchable database, so decisions, fixes, and dead ends from last month are one query away instead of being re-derived from scratch. The make-plan/do pair is a capable planning-execution workflow on its own.

Try it

claude

> /claude-mem:mem-search how did we fix the webhook race condition?