claude-supermemory plugin
The Supermemory plugin for Claude Code: persistent memory across sessions with automatic prompt recall, conversation capture and a codebase index
- 2,794 ★ on the repository
- Last commit
Broad access
Why this level
- Automatic capture sends the content of your agent conversations to the external Supermemory service
- The pre-save secret filter does not catch every token and password and does not check explicit MCP calls at all
- Hooks run at every session start, prompt and stop, launching Node scripts on your machine
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/claude-supermemoryDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Automatic capture sends the content of your agent conversations to the external Supermemory service
- The pre-save secret filter does not catch every token and password and does not check explicit MCP calls at all
- Hooks run at every session start, prompt and stop, launching Node scripts on your machine
This is third-party code. Review the repository files before installing.
What it does
claude-supermemory is a plugin that adds persistent memory to Claude Code on top of the Supermemory service. It registers four hooks in lifecycle order. At session start, session-start loads the profile and shows a welcome-back notice if you have not opened the project in over six hours. On each substantive prompt, recall-directive searches Supermemory itself and injects fresh matches without the model calling anything. The recall-approve hook auto-allows MCP tools that only read. When a session stops, the capture hook saves the new part of the conversation in the background. The /supermemory:index command writes the architecture, conventions and how to run into the project container, and /supermemory:status shows sign-in, API and MCP reachability and the active container. The context-gatherer subagent runs several searches across the project's containers and returns a brief. Codex and OpenCode use the same container tag, so repository memory is shared with them.
Who it is for. For developers who run long projects in Claude Code and are tired of re-explaining the same decisions to the agent every session.
Good fit when
- A project runs for months and the agent forgets decisions and their reasons each time
- You want the agent to pull past context on the topic of a request by itself, without being asked
- You work on one repository in Claude Code, Codex and OpenCode and want shared memory
- You want to describe the project to memory once so new sessions start with an understanding of its structure
- A team works on the project and needs a shared memory container configured in `.claude/.supermemory-claude/config.json`
Not a fit when
- The project holds secrets and private code and conversations must not go to an external service
- You need fully local memory without the cloud: the plugin works with the Supermemory service
- You only need instructions for the agent: a CLAUDE.md file is enough
- An old Claude Code version: the recall strip works from 2.1.287, and on 2.1.250 plugin validation rejects one of the events
Example request
Run /supermemory:index for this repository, then find what we decided about the auth scheme last weekLimitations
Needs Node.js 18 or later and a Supermemory API key, set through SUPERMEMORY_CC_API_KEY or obtained by browser sign-in. The plugin was renamed from claude-supermemory to supermemory and does not update in place, so it needs a fresh `/plugin install supermemory@supermemory-plugins`. The pre-save filter masks common tokens, passwords and private keys but cannot detect every secret. It does not rewrite memories already stored or local transcripts. If Supermemory is unreachable or the key is missing, the hooks exit cleanly and do not break the session.
How to disable. Run `/plugin uninstall supermemory@supermemory-plugins` in Claude Code; for the old version use `claude-supermemory@supermemory-plugins`. Settings live in ~/.supermemory-claude/settings.json and in the project `.claude/.supermemory-claude/config.json` and can be deleted by hand. Stored memories remain in Supermemory.
FAQ
How do I install Supermemory in Claude Code?
Run `/plugin marketplace add supermemoryai/claude-supermemory`, then `/plugin install supermemory`. Set SUPERMEMORY_CC_API_KEY from console.supermemory.ai or sign in through the browser on the first session. Node.js 18 or later is required.
What does the plugin send to Supermemory?
Automatic capture sends the new part of the conversation and, when configured, tool content. Before saving, an offline filter replaces secret-looking values with [REDACTED] markers but cannot catch every secret. Explicit MCP calls and recall queries do not pass through this filter.
How does memory recall work without a command?
When you are signed in, the prompt hook searches Supermemory itself on any substantive prompt and injects fresh matches before Claude sees it, with no permission prompt or MCP tool call. The recallDirective setting switches the mode: the hook stops searching and only tells the model when to search itself.
What do the plugin commands do?
There are two: project indexing and a status check. There is no separate search command; for deeper recall use the context-gatherer subagent or the MCP tools.
How do I migrate from the old claude-supermemory plugin?
The plugin was renamed to supermemory and will not update in place. Run `/plugin marketplace update supermemory-plugins` and `/plugin install supermemory@supermemory-plugins`, then uninstall the old one with `/plugin uninstall claude-supermemory@supermemory-plugins` if needed.
Discussion
No questions or reviews yet.
Related
PluginsA persistent memory plugin: stores observations from agent sessions and injects relevant context into new ones
Long-term memory for agents: a remote MCP server, plugins for Claude Code, Cursor and Codex, and an API for your own apps
A Claude Code skill that pulls memory from Memori Cloud before every non-trivial answer and writes new facts back to it
A plugin with memory of past chats: it searches by meaning for decisions and reasons across Claude Code, Codex, Cursor and opencode sessions over MCP
A self-improvement plugin for Claude Code, Codex and OpenCode: turns corrections into rules the agent follows in later sessions
Long-term agent memory in local Markdown files with a link graph, semantic search and MCP access
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| claude-supermemory (this entry) | yes | 2,794 | ||
| Supermemory | yes | OAuth | 31,369 | the general Supermemory MCP server and plugins for several agents, while this is the Claude Code plugin with hooks |
| Hindsight | yes | API key | 52,949 | a persistent memory system with its own server where the agent keeps experience and facts |
| CORE: memory and tasks for Claude Code | no | 2,003 | the CORE plugin for Claude Code keeps decisions and project context |
README in short
The description presents the plugin as persistent memory for Claude Code built on Supermemory. After installation come a feature table and a diagram of four hooks in lifecycle order. Then it covers shared agent memory through one container tag, environment variables, global and project settings, two commands and a privacy section with the secret filter.