search_savesLABSCO SUMMARY
Ask your own library a question instead of searching it.
Semantic search is the reason to connect this rather than open the app: "what have I saved on Postgres indexing" finds the right save without you remembering the title. Captures enrich asynchronously, so a fresh link comes back pending and is worth re-reading once it says ready.
What it is
A local stdio MCP server for Memlane, a personal library of saved links and notes. Your host spawns it with npx and it talks to the Memlane API with your key; every tool returns pretty-printed JSON in a single text block, with no streaming.
What you get
- search_saves — semantic search across embedded link and note content, each match scored 0–1
- list_saves — browse newest first with cursor pagination and filters for project, topic name, kind, status and a title substring
- get_save — one save's metadata, or the full markdown body with
includeContent: true - capture_url — save a link; extraction, summary, embedding and topic classification run in the background, so it returns
pendingand becomesreadylater - create_note / update_note — write and edit markdown notes, embedded immediately so they are searchable straight away
- list_projects / list_topics — resolve project and topic names to ids
- add_topic_to_save / remove_topic_from_save — tag and untag
- Deleting a save is not exposed as a tool; that stays on the REST API
Requirements
A Memlane account and an API key from Manage → Connections → API keys, passed as MEMLANE_API_KEY; MEMLANE_API_URL is optional and defaults to https://api.memlane.io. Node 18+ for npx. Keys carry read, write and capture scopes, and a call outside its scope fails with insufficient scope. ChatGPT, Claude Desktop and other remote clients cannot run npx — they use the hosted endpoint at https://api.memlane.io/mcp over OAuth, with no API key at all.
Setup effort
One command plus a key — npx -y @memlane/mcp, then supply credentials
USE CASES
01
Answering a question from things you already read
Search by meaning, then pull the best match's full markdown body.
02
Saving something mid-conversation
Capture the URL under a project, or write the thought straight in as a note.
03
Working through one project's material
Resolve the project id, then list its saves filtered by kind or status.
04
Tidying up how a save is filed
Find the topic id and attach or detach it.
10 tools
Listed in project docs. Your client may show a different list.
list_savesget_savecapture_urlcreate_noteupdate_notelist_projectslist_topicsadd_topic_to_saveremove_topic_from_saveBEFORE YOU INSTALL
What to know first
Tools you get
10 tools named in the project's documentation.
Requires
- npx on PATH
- MEMLANE_API_KEY in your environment
Where it runs
All 5 documented clients run the server on your machine.
INSTALLATION
Connect your client
Server setup
npx
TERMINAL
npx -y @memlane/mcpHosted endpoint
ENDPOINT URL
https://api.memlane.io/mcpAdd this hosted endpoint to your MCP client using the credentials listed under Requirements.
MCP Clients
Cursor, Windsurf, Gemini CLI
JSON
{
"mcpServers": {
"memlane": {
"command": "npx",
"args": [
"-y",
"@memlane/mcp"
],
"env": {
"MEMLANE_API_KEY": "ml_live_…"
}
}
}
}Claude Code
CLAUDE CODE · TERMINAL
claude mcp add memlane -- npx -y @memlane/mcpVS Code
VS CODE · JSON
{
"servers": {
"memlane": {
"command": "npx",
"args": [
"-y",
"@memlane/mcp"
],
"env": {
"MEMLANE_API_KEY": "ml_live_…"
}
}
}
}
