Remind CLI
Long-term memory layer for AI agents: decisions and facts are stored as JSON next to the code and found again in a later session.
- 84 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Writes memory files to `.remind/` and runs a local server when MCP is connected.
- The agent decides what to record, so secrets can reach memory that sits in the repository.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/remindDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Writes memory files to `.remind/` and runs a local server when MCP is connected.
- The agent decides what to record, so secrets can reach memory that sits in the repository.
This is third-party code. Review the repository files before installing.
What it does
Remind keeps an agent's experience across sessions. The agent records episodes (decisions, outcomes of attempts, preferences, facts), generalizes them into concepts and later searches by meaning. Remind itself makes no model calls: the agent does the thinking and the store stays deterministic. Memory lives in JSON files under `.remind/memory/` and travels with the repository. SQLite, PostgreSQL or MySQL can replace the files. Facts carry validity windows, and on every write Remind shows the nearest episodes and possible contradictions. Use the `remind remember`, `remind recall`, `remind snapshot` and `remind apply` commands or the `remind-mcp` MCP server.
Who it is for. Developers who want Claude Code, Cursor or Codex to remember project decisions between sessions.
Good fit when
- An agent keeps proposing an approach that already failed, and the outcome of the last attempt must be kept.
- A choice of database or library should live next to the code and land in commits.
- Several agents work on one project and read shared memory over MCP.
- A new session should start from `remind snapshot bootstrap` instead of a manual project recap.
Not a fit when
- A `CLAUDE.md` or `AGENTS.md` file with a few rules is enough.
- Memory cannot be stored in the repository: `.remind/memory/` files reach commits unless excluded.
- You need automatic summarizing without an agent: Remind does not generalize memory by itself.
Example request
Remember that we chose PostgreSQL for JSONB and remind me in the next session before I change the schema.Limitations
Python 3.11 or newer is required. Local embeddings need no key, while OpenAI embeddings and reranking come as separate packages. Memory does not generalize itself: quality depends on what the agent records and how it condenses it into concepts. Secrets from a conversation can end up in memory, and the files sit in the repository, so review them before committing. The storage format has changed before: read the migration page before upgrading. The MCP server in the example listens on a local address.
How to disable. Remove `remind` from the MCP configuration and the skills from `.agents/skills` or `.claude/skills`. `pip uninstall remind-mcp` removes the package, and delete the `.remind/` folder if you no longer need the memory.
FAQ
How do I install Remind for Claude Code?
Run `pip install remind-mcp`, then `remind skill-install --agent claude`: the three skills go to `.claude/skills`. Without the flag they land in `.agents/skills`.
Is an API key required?
No. Embeddings are computed locally with fastembed by default. OpenAI embeddings come from the separate `remind-mcp[openai]` extra.
How do I connect the MCP server?
Run `remind-mcp --port 8765` and point the client to `http://127.0.0.1:8765/sse?db=my-project`. The `db` parameter selects the store.
Where is the memory stored?
In `.remind/memory/` as JSON files you can commit. Indexes and logs live in `.remind/cache/` and `.remind/runtime/` and stay out of commits.
How is Remind different from a CLAUDE.md file?
`CLAUDE.md` is static and edited by hand. Remind stores episodes, time-bound facts and links between concepts, and the agent searches them by meaning.
Discussion
No questions or reviews yet.
Related
CLI toolsMCP server with a memory of the agent's own: it decides what to keep and wakes with it each session, stored in one SQLite file
A long-term memory MCP server for AI agents: facts fade along a forgetting curve and related records are compressed into a summary
Long-term memory for AI agents: Markdown files as the source of truth, a SQLite index as a cache, shared by Claude Code, Codex and others
Local memory for agents with a review queue: SQLite, an audit log and a token ledger, one vault for Claude Code, Codex and Cursor
A long-term memory system for AI agents: the agent stores experience and facts, then retrieves them through a built-in MCP server
Long-term memory for AI agents: a skill and the mindmemos command store facts from conversations and recall them in later sessions
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Remind (this card) | no | 84 | ||
| Basic Memory | no | not required | 4,566 | stores notes as Markdown over MCP; Remind records episodes and time-bound facts as JSON and flags contradictions on write |
| MCP Memory Service | no | not required | 2,117 | a standalone memory service; Remind works as a CLI with repository storage and ready-made skills |
README in short
The quick start reduces to one install and four commands. Next come storage, the batch commands `snapshot` and `apply`, the `remind events` change feed, `remind-eval` evaluation suites and the MCP server.