Memofs MCP server
{
"mcpServers": {
"memofs": {
"command": "npx",
"args": [
"-y",
"@memofs/mcp-server@latest"
]
}
}
}{
"mcpServers": {
"memofs": {
"command": "npx",
"args": [
"-y",
"@memofs/mcp-server@latest"
]
}
}
}{
"mcpServers": {
"memofs": {
"command": "npx",
"args": [
"-y",
"@memofs/mcp-server@latest"
]
}
}
}{
"servers": {
"memofs": {
"command": "npx",
"args": [
"-y",
"@memofs/mcp-server@latest"
]
}
}
}{
"mcpServers": {
"memofs": {
"command": "npx",
"args": [
"-y",
"@memofs/mcp-server@latest"
]
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y @memofs/mcp-serverversion 1.3.0-beta.3
What it does
MemoFS is a file-first memory runtime for AI agents that stores memory as plain Markdown and JSONL files in a `.memofs/` directory. It provides local-first persistent memory with optional cloud sync, designed for both end-users running AI agents and developers building custom agent systems.
- Initialize and manage durable agent memory across sessions using the CLI without manual configuration.
- Embed persistent memory into TypeScript or Node.js agent applications with read, write, and recall methods.
- Store and retrieve agent memories as inspectable, versionable files that can be committed to git and audited.
Tools this server exposes
| Tool | Does |
|---|---|
| sync | Synchronizes local memory with cloud storage using configurable read/write policies. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.