Long-term memory for coding agents that stores memories in a local SQLite database and shares them across Claude Code, Codex, Cursor and other MCP clients.
790+2 this week47TypeScriptMITWell maintainedUpdated 11 hours ago
qualixar/superlocalmemory
Local-first, governed memory layer that gives AI agents like Claude Code, Cursor and Codex persistent recall that admits when it doesn't know.
- Stars
- 273
- Forks
- 38
- Open issues
- 6
- Last commit
- 10 days ago
What is it?
SuperLocalMemory is a memory service for AI agents that runs on your own machine. It stores decisions, notes, documents and screenshots, then returns them through hybrid search when an agent asks. Recall answers come back in tiers: a supported answer, the closest memories clearly labelled as partial matches, or a plain statement that nothing in memory answers the question, so agents are less likely to build on a wrong guess.
It connects to agents through MCP or the command line. Plugins are provided for Claude Code, Codex, Cursor, Windsurf, VS Code Copilot and several other hosts. The core remember and recall path in its default mode needs no API key or cloud account, while optional features such as an online answer check or cloud embeddings send data out only when enabled.
The project also includes governance features: roles and workspaces, attributed writes, a hash-chained audit trail, export and erasure for GDPR-style requests, and a shared mesh that lets agents on one or several machines exchange messages and locks.
Key features
- 01Hybrid search combining semantic, keyword and entity-graph channels, with per-channel scores via a trace command.
- 02Three-tier recall that says when the answer is not in memory.
- 03Works with MCP, CLI and REST, with ready-made plugins for many coding agents.
- 04Local document and picture search, including PDFs indexed page by page.
- 05Roles, workspaces, audit trail and GDPR-style export and erasure.
- 06Mesh for messaging and locks between agents on one or several machines.
Use cases
- Giving Claude Code or Cursor a persistent memory of project decisions across sessions.
- Keeping a private, searchable store of team notes and documents that never leaves the machine.
- Letting several coding agents share context and avoid editing the same file simultaneously.
- Adding local memory to a custom agent built with LangChain, LlamaIndex or CrewAI.
Quick start
- 1
Install SuperLocalMemory globally with npm (Node 18+)
npm install -g superlocalmemory - 2
Run the setup and choose Mode A to keep everything on this machine
slm setup - 3
Connect SuperLocalMemory to your AI agent (for example Claude Code)
slm connect claude-code
Things to know
- Licensed under AGPL-3.0, a copyleft license that requires sharing modified source code when the software is offered over a network.
- The optional pictures and documents feature needs about 1.5 GB of models and 16 GB of memory.
- The on-device Laya checker is limited to Apple Silicon Macs; the online alternative requires opting in to send data out.
In collections
Topics
FAQ
- What is superlocalmemory?
- Local-first, governed memory layer that gives AI agents like Claude Code, Cursor and Codex persistent recall that admits when it doesn't know. SuperLocalMemory is a memory service for AI agents that runs on your own machine. It stores decisions, notes, documents and screenshots, then returns them through hybrid search when an agent asks. Recall answers come back in tiers: a supported answer, the closest memories clearly labelled as partial matches, or a plain statement that nothing in memory answers the question, so agents are less likely to build on a wrong guess.
- Is superlocalmemory free to use? What license does it have?
- superlocalmemory is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
- Is superlocalmemory still maintained?
- Actively maintained. The last commit to superlocalmemory was on Oct 28, 2026.
- What can I use superlocalmemory for?
- Giving Claude Code or Cursor a persistent memory of project decisions across sessions. Keeping a private, searchable store of team notes and documents that never leaves the machine. Letting several coding agents share context and avoid editing the same file simultaneously. Adding local memory to a custom agent built with LangChain, LlamaIndex or CrewAI.
- How do I install superlocalmemory?
- Install SuperLocalMemory globally with npm (Node 18+): npm install -g superlocalmemory Run the setup and choose Mode A to keep everything on this machine: slm setup Connect SuperLocalMemory to your AI agent (for example Claude Code): slm connect claude-code
- Where can I get superlocalmemory?
- The source code and installation instructions are on GitHub: https://github.com/qualixar/superlocalmemory
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A local MCP memory server that records every AI coding session word-for-word and feeds relevant past context back to agents like Claude Code and Cursor.
921+4 this week111PythonMITUpdated 6 days ago
Local, single-binary persistent memory for AI agents, using a four-graph knowledge store and LLM-supervised recall and writes across many coding agent runtimes.
698+21 this week80GoApache-2.0Well maintainedUpdated 7 days ago
Local-first persistent memory for AI agents, stored as markdown files with a knowledge graph, retrieval, and an MCP server.
33532TypeScriptApache-2.0Well maintainedUpdated last month
Persistent memory for AI coding agents that captures sessions automatically and injects relevant context into later sessions across Claude Code, Cursor, Codex and other MCP clients.
30.2K+141 this week2.7KTypeScriptApache-2.0Updated 4 hours ago
Long-term memory runtime for AI agents that stores memories as plain Markdown, ranks recall locally, and shares one store across Claude Code, Codex CLI and other shell-capable agents.
4.3K204PythonMITWell maintainedUpdated yesterday