Memory MCP Servers

Remnant by Dedale-Project lets agents search public debugging experiences and inspect provenance without an account, then contribute through a separate authenticated connection.

Persistent memory for your AI agents. Memory-MCP stores, retrieves, and semantically searches your AI agents' "memories", organized into isolated multi-tenant spaces protected by API Key — so context survives across conversations, without wasting tokens.

An open knowledge commons for AI agents: search sourced, versioned how-tos, read the exact revision, and report whether it worked.

Local-first persistent memory for AI coding agents. Markdown + SQLite hybrid retrieval (FTS5 + embeddings), works offline, no account. LongMemEval R@5 96.0%, zero LLM calls on retrieval.

One private, local memory shared by Claude Code and Codex. Single binary, no network, tamper-evident journal.

Private memory, documents, voice and a local LLM for your coding agents, served on-device by the M1K3 Mac app. 18 tools over a token-protected loopback HTTP server. Source-available.

Remembers why you approved, edited or rejected your agent's drafts and adds the closest past corrections before the next draft. MCP server + Claude Code plugin, local Postgres + pgvector.

Eight MCP tools (search, load, save, pick, tag, tags, untag, get_stats) over a local SQLite FTS5 index of past Claude Code and Codex sessions.

Local, zero-dependency memory MCP for AI agents: agents stage, humans promote, every entry has a receipt.

Gives Claude or ChatGPT the athlete context it needs to coach you properly: goals, PBs, thresholds, races, injuries and constraints for running, cycling and triathlon. Set up your profile once; the assistant reads it in every conversation and updates it when you ask. Hosted connector with OAuth sign-in. Stores no activity data.

Local-first memory of how you write: style profile and real reply examples from your own chat exports.

Self-hosted persistent cognitive memory substrate for AI coding agents with zero-amnesia context layer, blast-radius traversal, and opt-in libSQL serverless storage.

The team's shared memory that reaches an AI before it starts work.

Q&A network where AI agents ask, answer, vote and earn reputation — humans browse.

AI 도구를 위한 공유 메모리. 결정을 한 번만 저장하면 Claude, ChatGPT, Codex, Gemini 및 모든 MCP 클라이언트에서 프로젝트 및 저장된 기억을 불러올 수 있습니다. 원본 텍스트가 그대로 유지되며, OAuth 로그인이 지원되고, 최대 1,000개의 메모리를 무료로 사용할 수 있습니다.

One wiki for all your agents: pages read and written over MCP, a link graph, and broken-link checks.

Share context between agents and humans

A shared knowledge base people and AI agents read and write together over MCP. Free instances; public workspaces are free.

Agent-native spaced recall — deterministic system of record for Muse/AI agents. Remote MCP at fluidrecall.com/mcp; Markdown and JSON APIs also first-class.

Brainfile memory layer for Claude Code over MCP: free starter files + weekly freshness feed.

Local-first AI memory MCP server. Lets any MCP-compatible client read/write persistent memory (record / recall / consolidation) on the user's own machine. Privacy-first: nothing leaves the device.

Keep your memories and notes available across Claude, ChatGPT, and coding tools through a hosted MCP server.

Developer-alpha MCP server maintained by Louis Beaumont, with search and retrieval over encrypted, append-only knowledge through scoped, expiring grants.

Governed, portable memory for AI agents: every recall returns who asserted a fact, since when it has been true, and what replaced it. Facts are never overwritten. Local SQLite, no API key.

Tamper-evident, local-first memory for AI agents — hash-chained ledger, Apache-2.0, ships the Claude Code Seatbelt.

A self-contained markdown vault with a built-in MCP server. Humans edit through a web UI, agents talk to it over MCP, everything lives in plain .md files

Scans a folder of Markdown docs and fixes missing Open Knowledge Format fields so AI agents can use them. Runs locally; nothing leaves your machine.

Local memory your AI tools share: decisions, notes and corrections that carry across sessions.

Managed long-term memory for AI agents. Facts evolve as information changes, and agents retrieve relevant context without a custom memory system.

Persistent semantic memory shared across Claude, ChatGPT and all your AI assistants. Encrypted, EU-hosted, private by design.