Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@@memlane/mcp search my saves for 'artificial intelligence'"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Connect Memlane, your personal library of saved links and notes, to AI assistants that support local MCP servers.
Related MCP server: memos-mcp-server
This server runs locally via stdio. Most compatible hosts use the same JSON shape:
{
"mcpServers": {
"memlane": {
"command": "npx",
"args": ["-y", "@memlane/mcp"],
"env": {
"MEMLANE_API_KEY": "ml_live_…"
}
}
}
}Variable | Required | Default |
| Yes | n/a |
| No |
|
After npm install -g @memlane/mcp, you can use "command": "memlane-mcp" instead of npx.
Host | Where to add the config |
Settings → MCP, or | |
| |
| |
VS Code + Copilot |
|
|
For ChatGPT and Claude Desktop, use the remote MCP endpoint instead (OAuth, no API key).
Restart the host app after editing config files.
VS Code uses "servers" as the root key (not "mcpServers"):
{
"servers": {
"memlane": {
"command": "npx",
"args": ["-y", "@memlane/mcp"],
"env": {
"MEMLANE_API_KEY": "ml_live_…"
}
}
}
}Enable MCP in VS Code settings: "chat.mcp.enabled": true.
ChatGPT, Claude Desktop, Claude on the web, and other remote MCP clients cannot run npx. Use Memlane's hosted endpoint instead:
URL: https://api.memlane.io/mcp
In your AI app, add a custom MCP connector with that URL.
Complete the OAuth sign-in flow (Memlane account: magic link or future social providers).
Approve access on the consent screen.
No API key required. Disconnect by removing the connector in your AI app's settings.
The local server exposes 10 MCP tools. Full function reference:
docs/TOOLS.md — what each tool does, parameters, return shapes, and workflow tips
docs/API.md — REST mappings, auth scopes, and curl examples
Tool | Description |
| Semantic search over your library |
| Browse saves with filters (project, topic, kind, status) |
| Fetch metadata; optional full markdown body |
| Save a URL (enrichment runs in the background) |
| Create or edit markdown notes |
| List projects and topics |
| Tag or untag a save |
API keys are secrets. Store them in your MCP host's environment config. Never commit them to version control. Revoke a leaked key in Manage → Connections → API keys.
MIT
Attach a topic to a saved link or note
Inputs: itemId (required), topicId (required).
Save a URL to Memlane (background enrichment: extract, embed, classify)
Inputs: url (required), projectId.
Create a markdown note in Memlane
Inputs: title (required), content (required), projectId.
Get a saved link or note by id, optionally including full markdown content
Inputs: id (required), includeContent.
List Memlane projects
Inputs: no named parameters.
List saved links and notes, newest first
Inputs: limit, cursor, projectId, topicName, +3 more.
List Memlane topics (general and project-scoped)
Inputs: no named parameters.
Remove a topic from a saved link or note
Inputs: itemId (required), topicId (required).
Semantic search over your saved links and notes in Memlane
Inputs: query (required), limit.
Update a note's title and/or markdown body
Inputs: id (required), title, content.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 10 tools
Each tool targets a distinct action (capture, create, update, get, list, search, add/remove topic) on distinct resources (saves, notes, topics, projects). No overlap or ambiguity.
All tool names follow a consistent verb_noun pattern in snake_case (e.g., capture_url, list_saves, remove_topic_from_save). No mixing of conventions.
With 10 tools covering core operations for save/note management, topic organization, and project listing, the count is well-scoped and appropriate for the server's purpose.
Notable gaps: no delete tool for saves/notes, no create or update tools for topics or projects. The surface covers basic CRUD for notes and saves but misses lifecycle operations.
Personal knowledge MCP: capture bookmarks, notes & todos by chat; archive pages; search memory.
Persistent personal memory for AI assistants — save, search, and recall across every MCP client.
Persistent memory for AI assistants. Save once; recall from Claude, ChatGPT, or any MCP client.
MCP-native notes and memory for ChatGPT, Claude, and other AI tools.
