what is cognee-mcp-server?
The cognee-mcp-server is a server designed for 'cognee', an AI memory engine that builds a knowledge graph from user-input text and facilitates searching within it.
how to use cognee-mcp-server?
To use the cognee-mcp-server, configure it according to your environment, including setting up necessary parameters in the claude_desktop_config.json, and run it with the specified commands related to your installation path.
key features of cognee-mcp-server?
- Constructs knowledge graphs based on textual input.
- Allows for searching within the generated knowledge graph.
- Supports custom graph model implementations through flexible configuration options.
use cases of cognee-mcp-server?
- Enhancing AI systems with memory by building contextual knowledge graphs.
- Performing efficient searches across complex datasets based on user queries.
- Supporting applications needing robust data retrieval and organization capabilities.
FAQ from cognee-mcp-server?
- What is required to run cognee-mcp-server?
You will need to configure the server with your API key and local paths, and it works with various database providers.
- Can I use custom models with cognee-mcp-server?
Yes, you can use your own Pydantic graph model implementations by specifying the filenames and class names in the configuration.
- Is cognee-mcp-server compatible with existing AI frameworks?
Yes, it is designed to integrate smoothly with systems like Claude Desktop and can be adapted for other AI frameworks.
cognee-mcp-server
An MCP server for cognee, an AI memory engine.
Tools
Cognify_and_search: Builds knowledge graph from the input text and performs search in it.- Inputs:
text(String): Context for knowledge graph contstructionsearch_query(String): Query for retrievalgraph_model_file(String, optional): Filename of a custom pydantic graph model implementationgraph_model_name(String, optional): Class name of a custom pydantic graph model implementation
- Output:
- Retrieved edges of the knowledge graph
- Inputs:
Configuration
Usage with Claude Desktop
Add this to your claude_desktop_config.json:
Using uvx
"mcpcognee": {
"command": "uv",
"args": [
"--directory",
"/path/to/your/cognee-mcp-server",
"run",
"mcpcognee"
],
"env": {
"ENV": "local",
"TOKENIZERS_PARALLELISM": "false",
"LLM_API_KEY": “your llm api key”,
"GRAPH_DATABASE_PROVIDER": “networkx”,
"VECTOR_DB_PROVIDER": "lancedb",
"DB_PROVIDER": "sqlite",
"DB_NAME": “cognee_db”
}
}
