Letta MCP Server
by oculairmedia
The Letta MCP Server provides tools for agent management, memory operations, and integration with the Letta system. It allows users to manage agents, memory blocks, and tools within the Letta ecosystem.
Last updated: N/A
What is Letta MCP Server?
The Letta MCP Server is a backend server designed to facilitate the management of agents, memory, and tools within the Letta ecosystem. It provides a set of APIs and tools for creating, managing, and interacting with agents and their associated resources.
How to use Letta MCP Server?
The server can be run using Node.js or Docker. Configuration is done via environment variables and can be integrated with Letta's mcp_settings.json. The server exposes endpoints for various tools, which can be accessed via HTTP requests using SSE or other transport mechanisms.
Key features of Letta MCP Server
Agent Management (create, list, prompt, get, modify, delete, clone, bulk delete)
Memory Management (list, create, read, update, attach)
Tool Management (list, attach, upload, bulk attach)
Model Management (list LLM and embedding models)
Archive Management (list, create, modify, delete passages)
MCP Server Management (list MCP servers and tools)
Import/Export Agents
Docker support for easy deployment
Use cases of Letta MCP Server
Managing AI agents within the Letta framework.
Storing and retrieving information for agents using memory blocks.
Attaching and managing tools for agents to perform specific tasks.
Integrating with the Letta system for seamless agent interaction.
Automating agent creation and management workflows.
Extending agent capabilities with custom tools.
FAQ from Letta MCP Server
How do I create a new agent?
How do I create a new agent?
Use the create_agent tool with the required parameters: name and description. Optional parameters include model and embedding.
How do I list all available agents?
How do I list all available agents?
Use the list_agents tool. You can optionally filter the results.
How do I attach a tool to an agent?
How do I attach a tool to an agent?
Use the attach_tool tool, providing the agent_id and either tool_id, tool_ids, or tool_names.
How do I run the server using Docker?
How do I run the server using Docker?
Use the provided Docker commands to build and run the image. Ensure you set the necessary environment variables like PORT and NODE_ENV.
How do I configure the server with Letta?
How do I configure the server with Letta?
Add the server configuration to your mcp_settings.json file, specifying the command, arguments, environment variables, and allowed tools.