memtomem — Claude Code MCP server: memory and setup
A local memory layer for AI coding agents that indexes notes and code, then retrieves them with keyword and semantic search.
claude mcp add memtomem -- uvx memtomemWhat memtomem does
memtomem builds a searchable knowledge base from Markdown notes, documents, and code. It handles .md, .json, .yaml, .toml, .py, and JavaScript/TypeScript files, combining BM25 keyword matching with vector similarity. Your files remain the source of truth while the database can be rebuilt.
The server runs locally and connects to Claude Code with claude mcp add memtomem -- uvx memtomem. Core memory actions happen when the agent calls them; client hooks are optional integrations. Python 3.12 or newer is required, so check that runtime before setup.
What it helps with
- carry notes across agent sessions
- retrieve related content by meaning
- index Markdown and code together
- give Claude Code access to local memory
Example prompts for the agent
- Check the memtomem connection and call the
mem_statustool. - Search memory for deployment notes and return related entries.
- Save this decision to memory with the
opstag.
How to install memtomem
Platform CLI
claude mcp add memtomem -- uvx memtomemuvx runs the command named after the package; if the README names another one, use uvx --from memtomem <command>
Platform CLI
codex mcp add memtomem -- uvx memtomemuvx runs the command named after the package; if the README names another one, use uvx --from memtomem <command>
Config file
{
"mcpServers": {
"memtomem": {
"command": "uvx",
"args": [
"memtomem"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from memtomem <command>
Config file
{
"servers": {
"memtomem": {
"type": "stdio",
"command": "uvx",
"args": [
"memtomem"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from memtomem <command>
Platform CLI
gemini mcp add memtomem uvx memtomemuvx runs the command named after the package; if the README names another one, use uvx --from memtomem <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about memtomem
How do I install memtomem in Claude Code?
Run claude mcp add memtomem -- uvx memtomem.
What does memtomem do?
It indexes notes, documents, and code, then retrieves them through keyword matching and semantic similarity. An AI agent accesses the local index through MCP.
What does memtomem require?
Python 3.12 or newer is required. The server runs locally, and Markdown files remain the source of the stored knowledge.
