Frequently Asked Questions (FAQ) about Picard MCP Server

Q: What is the Picard MCP Server? A: The Picard MCP Server is a memory management system built on the Model Context Protocol (MCP), designed to provide secure and semantically searchable memory for Large Language Models (LLMs).

Q: What is the Model Context Protocol (MCP)? A: MCP is an open standard that defines how applications provide context to LLMs, acting as a bridge to external data and tools.

Q: What are the key features of the Picard MCP Server? A: Key features include OAuth 2.0 authentication, PostgreSQL memory storage with pgvector, permission-based access control, vector embeddings for semantic search, and LLM integration.

Q: How does the Picard MCP Server ensure data security? A: The server uses OAuth 2.0 with PKCE for authentication, encrypts memory text content at rest, and employs permission-based access controls.

Q: What is the Django Client and what is its role? A: The Django Client is a web application demonstrating how to integrate with the MCP server, providing a user interface for memory management and querying.

Q: What are the two main authentication approaches supported by the system? A: The system supports Direct Connect with User Context Token Flow (recommended) and the standard OAuth 2.0 Authorization Code Flow with PKCE (legacy).

Q: How are memories stored and managed in the system? A: Memories are stored as text with associated metadata, including vector embeddings for semantic search. Permissions control who can access each memory.

Q: What is semantic search and how does the Picard MCP Server implement it? A: Semantic search allows finding memories based on meaning using vector embeddings. The server uses OpenAI’s text-embedding-3-small model for generating embeddings and pgvector for efficient vector storage and retrieval.

Q: What is the purpose of UUIDs in the system? A: UUIDs (Universally Unique Identifiers) are used instead of sequential integers for security, scalability, non-guessability, and consistency.

Q: How do I set up and deploy the Picard MCP Server? A: The setup involves using Docker and Docker Compose. You’ll need to clone the repository, configure environment files, start the services, create an admin user, and register the Django client.

Q: How can I test if the Picard MCP Server is working correctly? A: Run the MCP server and Django client tests. Also, perform manual testing by creating a user account, logging in, connecting to the MCP server, and creating/managing memories.

Q: Where can I find documentation for the Picard MCP Server API? A: The MCP server includes Swagger/OpenAPI documentation, accessible at /docs when the server is running.

Q: What is the license for the Picard MCP Server? A: The Picard MCP Server is licensed under the MIT License.

Q: How does the Picard MCP Server integrate with the UBOS Platform? A: Integrating Picard MCP Server into the UBOS platform empowers your AI Agents with long-term memory and contextual awareness, enabling them to perform more sophisticated tasks and provide more personalized experiences. It provides more security, more reliable and better AI Agents for your business.