All MCP Servers
AI & MemoryMCP Registry17 toolsB · 64

mcp-memory-rs

io.github.DioNanos · Updated

Local-first agent memory: versioned JSON categories, BM25 search, per-device ACL, fleet sync.

At a glance

mcp-memory-rs is an MCP server in the AI & Memory category. It publishes no package or endpoint yet. It exposes 17 tools, including memory_read, memory_write and memory_append. It scores B (64/100) on the MCP Skills Hub health score as of Oct 28, 2026.

Source 2GitHub stars

How to install mcp-memory-rs in your AI client

Setup for Claude Code, Claude Desktop, Cursor, Codex, Windsurf, Cline and VS Code — copy-paste ready.

How to install mcp-memory-rs in Claude CodeAnthropic's agentic CLI

  1. No installation available

    This registry entry publishes neither a package nor a remote endpoint, so there is nothing to install yet. Check the source repository for manual setup instructions.

What tools does mcp-memory-rs expose? (17)

Taken from the project's README — the server itself was not queried, so check its docs for the exact schemas.

  • memory_readRead a category (optional field filtering).
  • memory_writeReplace or merge-patch a category; versioned; `expected_hash` precondition.
  • memory_appendAppend a timestamped entry to a bounded append-only log category; auto-prunes by `max_entries`/`max_age_days`. For event streams / session journals, so they don't bloat memory categories.
  • memory_deleteDelete a category (backup created first).
  • memory_listAll categories with hash/size/last-update metadata.
  • memory_searchFTS5 full-text search, BM25 ranking, category/date/actor filters, snippets.
  • memory_search_semanticHybrid TF-IDF + FTS5 search.
  • memory_historyVersion history of a category.
  • memory_deltaChanges since a known hash (cheap polling).
  • memory_contextMulti-category warmup read, token-budget oriented.
  • memory_compactPrune old versions and backups.
  • memory_statusLocal-first status: dirty queue, manifest hash, last sync.
  • memory_doctorDiagnostics: paths, database, categories, sync config.
  • sync_manifestMerkle-rooted manifest of all categories.
  • sync_diffCompare local vs remote manifest: push/pull/conflict sets.
  • sync_pushExport/import sync envelopes.
  • memory_syncOne sync step against the configured remote (`push_dirty` / `pull_remote`).

What the code can do

This registry entry publishes neither a package nor an endpoint, so there is nothing to analyze.

Before you install

Signals gathered from the source repo and package registry — not a safety verdict.

  • Last commit 1 month ago — Actively maintained.
  • 2 GitHub stars — Very little community usage so far.

Client compatibility

Claude Code: UnavailableClaude Desktop: UnavailableCursor: UnavailableCodex CLI: UnavailableWindsurf: UnavailableCline: UnavailableVS Code: Unavailable

Health score

B · 64

Maintenance, adoption, transparency and packaging — not a safety audit. How this is calculated

  • Maintenance30%98%

    Last commit 43 days ago.

  • Adoption20%14%

    2 stars.

  • Transparency20%100%

    4/4: ✓ public source repo, ✓ declared license, ✓ documentation, ✓ discoverable tool list.

  • Provenance15%75%

    Links to its own repository; package linkage unverified.

  • Packaging15%0%

    Publishes neither a package nor an endpoint — nothing installable.

Documentation

Full README

mcp-memory-rs

Local-first MCP memory server in pure Rust. Persistent, versioned, queryable memory for AI agents — owned by the user, not by a platform.

State lives on your disk as plain JSON categories backed by SQLite (FTS5). Every write is versioned and backed up. A fleet of devices syncs through an optional HTTP endpoint with merkle manifests and optimistic concurrency. No cloud account, no embedding service, no network dependency in the default mode.

Why

Agent memory tied to a vendor dies with the vendor session. This server keeps the agent's long-lived state — identity, projects, infrastructure notes, per-device workflows — in files you can read, grep, diff, and back up yourself. Any MCP client (Claude Code, Codex, or anything speaking MCP stdio) gets the same memory; swapping the model does not lose the state.

Design choices that follow from that:

  • Local-first: the stdio server works fully offline against local storage.

Sync is a separate, explicit step — never a hidden dependency.

  • Plain storage: one JSON file per category plus a SQLite index. The

database can be rebuilt from the files; the files are the truth.

  • Versioning by default: every write keeps history; deletes create a

backup first. memory_history and memory_compact manage the tail.

  • Concurrency-safe writes: expected_hash turns clobbering races between

concurrent agents into explicit conflicts.

Excerpt — the full document is on GitHub.

Maintain mcp-memory-rs? Add a badge to your README

Shows its health grade, and links readers to install steps for every client.

markdown
[![mcp-memory-rs on MCP Skills Hub](https://mcp-skills-hub.vercel.app/badge/io-github-dionanos-mcp-memory-rs)](https://mcp-skills-hub.vercel.app/mcps/io-github-dionanos-mcp-memory-rs)

Catalog data refreshed . Something wrong? Report it.