MemCP - Memory Context Protocol for AI Agents
About
A memory MCP for IDE AI agent context support based on the mcp_server example by Zep AI
Details
- License
- MIT license
Explore
- Temporal Knowledge Graph: Lets your AI agents automatically build and query a temporal knowledge graph that evolves as new information is added.
- MCP Integration: Works with any MCP-compatible clients.
- Entity Extraction: Automatic entity and relationship extraction with coding specific entity types. Customization is on the roadmap.
- Flexible Configuration: Simple configuration via environment variables, TOML files, or CLI arguments
- Persistence: Stores knowledge in Neo4j for persistence across sessions
- Multiple Transports: Supports both SSE (HTTP) and stdio transports for integration with different clients
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
MemCP - Memory Context Protocol for AI AgentsCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
1. Python 3.10 or higher
2. Neo4j database (version 5.26 or later)
> Note: you can run it with docker if you do not want to install Neo4j locally.
3. OpenAI API key (required for embeddings)
4. Anthropic API key (optional, for Claude models)
uv sync
cp .env.example .env
Add the following to your Cursor plugin configuration:
{
"mcpServers": {
"MemCP": {
"transport": "sse",
"url": "http://localhost:8000/sse"
}
}
}
For containerized deployment:
MemCP exposes these MCP tools to your LLM:
- add_episode: Add an episode to the knowledge graph (text, JSON, messages)
- search_nodes: Search for entity nodes in the graph
- search_facts: Search for relationships between entities
- delete_entity_edge: Delete a relationship between entities
- delete_episode: Delete an episode from the knowledge graph
- get_entity_edge: Get details about a specific relationship
- get_episodes: Retrieve recent episodes
- clear_graph: Reset the knowledge graph (use with caution)
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"memcp - memory context protocol for ai agents": {
"memcp": {
"command": "uv",
"args": [
"sync"
]
}
}
}
}McpServers
{
"memcp": {
"command": "uv",
"args": [
"sync"
]
}
}MemCP is an extensible memory MCP server for AI coding agents. It modularizes and extends the Zep AI Graphiti mcp-server example, allowing AI agents to build, maintain, and query a temporal knowledge graph of facts, entities, and relationships.
MemCP is currently designed as a plug and play memory server for any IDE or LLM client app that supports MCPs through local sse or stdio connections.
> Note: It has largely been tested with the Cursor IDE compared to other MCP capable clients.
Features
- Temporal Knowledge Graph: Lets your AI agents automatically build and query a temporal knowledge graph that evolves as new information is added.
- MCP Integration: Works with any MCP-compatible clients.
- Entity Extraction: Automatic entity and relationship extraction with coding specific entity types. Customization is on the roadmap.
- Flexible Configuration: Simple configuration via environment variables, TOML files, or CLI arguments
- Persistence: Stores knowledge in Neo4j for persistence across sessions
- Multiple Transports: Supports both SSE (HTTP) and stdio transports for integration with different clients
Documentation
Comprehensive documentation for MemCP can be found in the docs directory. See the docs README for an overview of available documentation.
Installation
Prerequisites
1. Python 3.10 or higher
2. Neo4j database (version 5.26 or later)
> Note: you can run it with docker if you do not want to install Neo4j locally.
3. OpenAI API key (required for embeddings)
4. Anthropic API key (optional, for Claude models)
Clone the repository
git clone https://github.com/evanmschultz/memcp.git
cd memcp
Using UV (Recommended)
UV is a fast package manager for Python written in Rust. Follow their docs for installation instructions if you don't have it already.
```bash
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
