Memory & Knowledge MCP Servers for Devin Desktop

Your Obsidian vault, connected to AI. MCP server for bridging personal knowledge with LLMs.

Persistent developer memory MCP server — the LLM is stateless, Mnemos isn't

Shared, persistent memory for AI agents, with semantic recall and live updates.

Per-project knowledge brain with auto-learning from AI conversations. MCP server + SQLite + vector search.

MCP server for Obsidian Companion

OneBrain MCP server — persistent AI memory for Claude, Cursor, and other MCP clients

MCP server for Claude Code and Claude Desktop to interact with Open WebUI — manage chats, RAG knowledge bases, functions, and more.

Org-wide persisted memory for AI coding agents via AWS Bedrock AgentCore. MCP server with short-term session events, long-term semantic

PAMPA – Protocol for Augmented Memory of Project Artifacts (MCP compatible)

Governed, self-hosted memory for AI agents: writes queue until a named person approves.

MCP server for hierarchical agent memory

Local-first memory layer for coding agents with hybrid keyword and semantic search.

MCP server for interacting with the Phiacta knowledge platform

Review gate for shared agent memory: an unreviewed or hand-edited memory is never served.

MCP server and Claude Code skill for project-aware code changes with low token usage.

Agent-native knowledge OS on Markdown: typed graph, hybrid search, and compiler over MCP.

MCP server for Raggy - universal memory for AI agents. Ships with a built-in agent protocol so every connected client auto-loads the rules

Local-first RAG indexing and semantic search MCP server

Local-first, cross-agent memory for AI coding agents (Claude, Codex, Gemini, Cursor, …) — one SQLite file, MCP over stdio or HTTP, FTS5

Reflect.app MCP connector (Model Context Protocol server).

Remembry MCP Server — manage your meeting knowledge base via AI assistants. Recording requires the Remembry desktop app to be running.