GPT OpenMemory

Provides local memory storage, search, and markdown export for ChatGPT Desktop via an MCP endpoint.

python

0

GitHub Stars

python

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "seanshin0214-openmemory-mcp": {
      "command": "python",
      "args": [
        "server.py"
      ]
    }
  }
}

GPT OpenMemory MCP Server lets you run a local MCP endpoint that stores, searches, and exports ChatGPT Desktop conversation memories. It replaces the built-in memory when you want persistent, local control over your dialogue history and related documents, all accessible via a lightweight MCP interface on your machine.

How to use

You connect an MCP client to the local server running on your machine. Use the client to save conversations, recall summaries, search memories by keywords, export conversations to Markdown, and store documents locally.

Typical workflows include saving a memorable chat after a productive session, asking for a structured memory abstract, or exporting a conversation for sharing. You can also store arbitrary documents and retrieve recent memories quickly.

How to install

Prerequisites: you need Python installed on your system.

pip install fastapi uvicorn aiosqlite aiofiles

Run the server with Python. The MCP server will start locally and listen on port 8769.

python server.py

After you start it, the MCP server will be available at http://127.0.0.1:8769. Use an MCP client to connect to that URL and begin interacting with memory features.

Notes and usage details

Data is stored in a local SQLite database named memory.db. Exported files can be saved to a user-specified path. You can perform memory-related actions via the MCP endpoints exposed by the server.

Available memory tools include saving memory, retrieving abstracts, updating abstracts, listing recent memories, performing searches, obtaining statistics, exporting to Markdown, saving documents, and listing exported files.

Security and maintenance

Since the server runs locally, ensure your machine is secured and not exposed to untrusted networks. If you enable remote access, protect it with appropriate authentication and network restrictions.

Monitor the memory database size and exported files to manage storage usage over time.

Available tools

save_memory

Stores a conversation memory into the local SQLite database for later recall.

recall_memory_abstract

Retrieves an abstract or summary of a memory based on stored content.

update_memory_abstract

Updates the saved abstract for a memory entry to refine or correct the summary.

get_recent_memories

Fetches a list of memories ordered by recency for quick review.

search_memories

Performs keyword and context-based searches across stored memories.

get_memory_stats

Returns statistics about stored memories, such as count and storage size.

export_to_markdown

Exports a memory or conversation thread to a Markdown (.md) file.

save_document

Saves arbitrary content as a local document file.

list_exported_files

Lists files that have been exported from memory collections.