Happy
Lets you run and check Claude Code and Codex sessions from your desktop and phone
by Happy Engineering
Claude-Mem
by Alex NewmanPluginWrite code
Lets Claude Code remember earlier sessions so you can pick up where you left off
The problem it solves
You open a new Claude Code session and spend the first ten minutes explaining the project all over again.
For Claude Code usersDevelopersKnowledge workers
- Before
- Every new session starts blank, so you repeat what was decided and what is already done.
- After
- New sessions open with short summaries of recent work, and you can search past sessions in plain words.
- Stars
- ★ 106k
- License
- Apache-2.0
- Cost
- Usage-based
- For
- Claude Code
- Contents
- Includes scripts
Overview
Claude-Mem is a memory plugin for Claude Code made by Alex Newman; it is not an Anthropic product. As you work, it records what the agent does, has an AI model compress those notes into short summaries and adds the relevant ones to the start of your next session. You can also search past work in plain words or browse it in a local web viewer. It suits people who use Claude Code every day on the same projects.
Each new session starts with a summary of earlier work, so you spend less time explaining a project to the agent again.
Try these first
Start with the first one: it is a small task you can check easily.
- What did we work on in this project in the last few sessions? List the decisions, and don't change any files.First try · read-only
- Search my memory for [topic] and show the related notes with their dates.
- Summarise what we did on [feature] last week and what was left unfinished.
- Keep this out of memory: <private>[paste text]</private>not stored
Example prompts written by Agent Flows Hub.
What you get
From the creator's documentation.
| Output | What it is | Reads |
|---|---|---|
| Session capture | Records what the agent does in each session through Claude Code hooks | Your computer |
| AI summaries | Compresses each session into short notes with your Claude plan, your own key or the hosted service | Model provider |
| Context at start | Adds relevant notes from recent sessions when a new session begins | Your computer |
| Memory search | Search past work in plain words, or browse it in a local web viewer | Your computer |
| Private tags | Text inside <private> tags is left out of storage | Your computer |
| Usage analytics | Anonymous usage counts are sent by default; one command turns them off | Internet |
Good to know
You work with Claude Code on the same project across many sessions. Claude-Mem saves short summaries of each session and brings the relevant ones into the next, so the agent starts with what was already done.
Limits
- Summaries are written by an AI model: your Claude plan, your own OpenRouter or Gemini key, or the hosted claude-mem service, which has a 30-day free trial and then needs a subscription or falls back to your Claude plan.
- Anonymous usage analytics are on by default; the docs give one command to turn them off. Wrap text in <private> tags to keep it out of storage.
Notes by Agent Flows Hub (original project name: claude-mem), written from the creator's documentation, not from running it ourselves. Claude-Mem installation guide · Claude-Mem README
Stuck setting it up?
Tell us what happened. We take a look first; if it needs paid work, you get an estimate before anything starts.
More to write code
Other plugins
- claude-obsidian Turns articles, transcripts and notes into a linked Obsidian wiki that cites its sources
- Codex for Claude Code Lets Claude Code ask OpenAI's Codex to review your changes or take over a task
- Design Research Skills Helps you plan user interviews and turn research notes into personas and journey maps
- Designer Toolkit Skills Helps you write design case studies, rationale, presentations and interface copy
- Discord for Claude Code Lets you message your running Claude Code session from Discord and get replies there
- Elephant-Goldfish Model Helps your coding agent draft a feature spec, then has fresh reviewers poke holes in it