MCP Memos Server

You can run this MCP Memos Server to give AI agents safe, read/write access to your Memos data through a consistent interface. The server supports creating, reading, updating, deleting, and searching memos, plus handy resources to access recent memos and by date. It can run locally with Python or be deployed via Docker, and you control access by providing a Memos API key at runtime.

How to use

You use the MCP Memos Server by starting the local MCP process and then connecting your MCP client to it. Start the server first, then supply your Memos API key using the set_api_key tool before performing any memo operations. Once connected, you can create memos, search content, retrieve specific memos by ID, update or delete memos, and fetch memos by date or date range. You can also access recent memos or direct memo content through the provided memo:// resources.

How to install

Prerequisites are installed on your machine and you have a running Memos server with an API key you can use at runtime.

# 1) Install the server using Docker
# If you prefer Docker, pull the image and run it with a provided env file
# 2) Or run locally with Python

Additional steps and usage notes

Configuration centers around pointing the MCP server at your Memos instance and controlling how memos are surfaced to AI agents. You provide the MEMOS_URL base address and, optionally, a MEMOS_API_KEY when you start or via the runtime prompt. The server’s default behavior is PRIVATE visibility for new memos unless you override it.

Configuration and security

Two main ways to run the MCP Memos Server are supported: via Docker and via a local Python run. In both cases you supply environment details so the server can connect to your Memos instance.

# Docker (recommended for quick start)
docker run --rm -i --env-file /path/to/your/.env memp-memos-server

# Local Python run
MEMOS_URL=https://your-memos-server.com
python server.py

Troubleshooting

If you have trouble connecting, verify MEMOS_URL is correct, ensure the API key (when provided) has the right permissions, and confirm the Memos server is reachable from the MCP host. If you’re using Docker, check container logs and ensure the env-file contains your MEMOS_URL and optional MEMOS_API_KEY. If you see permission errors, ensure the API key grants read/write access.

Tools and data access

The server exposes actions to manage memos and to access memo data with standardized endpoints. You can set your API key, create and modify memos, search memo content, retrieve memos by date or date range, and list recent memos.

Notes on running with Claude Desktop

You can configure Claude Desktop to connect to the MCP Memos Server. Use a stdio-based configuration that runs the MCP server either via Docker or locally with Python, and point Claude to the appropriate runtime command.

API usage and endpoints

Key operations you will perform through the MCP include setting the API key, creating memos, reading memos by ID, updating and deleting memos, and performing text-based searches. You also access memo data via resources such as recent memos, search results, memos by date, and individual memo content.

Available tools

set_api_key

Set your Memos API key for the current session. This must be called first before any other memo operations.

create_memo

Create a new memo with content, visibility, and tags.

get_memo

Retrieve a memo by its ID.

update_memo

Update an existing memo's content or settings.

delete_memo

Delete a memo from your Memos instance.

search_memos

Search memos by text content.

get_memos_by_date

Get memos created on a specific date.

get_memos_by_date_range

Get memos within a date range.

list_recent_memos

List your most recently created memos.