Memvid MCP Server

Example Use Cases

• 1

  Store conversational or contextual memory as compact, streamable media and retrieve relevant chunks by semantic similarity.

• 2

  A lightweight, local MCP-compatible endpoint to run semantic search over application-specific memories without a heavy vector DB setup.

• 3

  Integrate memvid-encoded memories into existing MCP clients using Streamable-HTTP for quick lookup and prototyping.

Description

This server uses memvid to encode text data into videos and exposes a Streamable-HTTP Model Context Protocol (MCP) interface for adding and searching memory chunks. It supports adding chunks (which resets the stored memory video) and semantic search queries (returns top-matching chunks by default). It's designed to let agents store and retrieve semantically searchable memory efficiently via a standard MCP endpoint. The project is open-source and runnable locally with minimal setup.