Show HN: Tessera – MCP server that gives Claude persistent memory and local RAG (github.com)

Tessera has been announced as a new MCP server designed to enhance Claude Desktop by enabling persistent memory and local retrieval-augmented generation (RAG). This tool addresses a significant limitation: while Claude Desktop can read and answer queries based on attached documents, it previously lacked the ability to search across an entire workspace efficiently. By indexing local documents such as Markdown, CSV, and session logs into a vector store, Tessera allows Claude to automatically search, respond with contextual insights, and retain memory of past interactions—all while ensuring complete data privacy by keeping everything on the user's machine. With no external dependencies required and a straightforward setup via pip installation, Tessera's features include cross-session memory, a knowledge graph visualizing document connections, and a hybrid semantic-keyword search. Technical integration involves a simple initialization process that configures document indexing and memory management, enabling users to track decisions and project status seamlessly. This advancement not only enhances productivity within organizations by effectively bridging information silos but also shifts the paradigm towards more intelligent and context-aware AI interactions in the workspace.