r3

io.github.n3wthPublished by GitHub account n3wthv1.3.2Updated Oct 24, 2026

Local Redis memory MCP for AI apps with hybrid caching and optional Mem0 cloud backup.

VerifiedSTDIODesktop onlyDatabasesKnowledge & Memory

Overview

AI-generated overview

Gives an assistant a local, persistent memory store with semantic search, backed by embedded Redis and a local vector index.

What you can do
r3 runs as a local MCP server that stores memories and retrieves them later, so preferences and facts survive across conversations. Tools include add_memory, search_memory, get_all_memories, get_memory, update_memory, delete_memory, deduplicate_memories, import_memories, plus cache_stats, sync_status and optimize_cache. An enhanced mode adds extract_entities, get_knowledge_graph and find_connections for entity and relationship lookups.
When to use it
Useful when you want an assistant to remember user preferences, notes or project context between sessions without setting up an external database or account. Also fits workflows that need semantic recall or a small local knowledge graph over stored memories.
Requirements
Runs locally over stdio, started with npx @n3wth/r3, so Node.js is needed. It downloads and manages an embedded Redis binary and keeps a local vector index on disk; no signup or API key is required to start. Optional environment variables: REDIS_URL to point at an external Redis, MEM0_API_KEY for cloud sync, MEM0_USER_ID for namespacing, INTELLIGENCE_MODE for enhanced or basic mode. Desktop clients only.
Before you install
Memory tools write, modify and delete stored data, and deduplicate_memories merges records, so review before letting an assistant clean up memories. Setting MEM0_API_KEY sends memories to Mem0's cloud API for cross-device sync; without it nothing leaves the machine. The README notes a native module build failure on some macOS setups.

Installation

In SourceWeft

  1. Open r3 in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

r3

[npm version] [npm downloads] [License: MIT]

Persistent memory for MCP clients (Claude Desktop, Claude Code, Cursor) that runs entirely on your machine, with no accounts, no API keys, and no cloud service required to start.

[Memory persisting across two separate server processes]

Quick start

bash
npx @n3wth/r3

That single command starts an MCP server backed by an embedded Redis instance and a local vector index. There is no separate database to install and no signup step.

How it works

MCP client (Claude Desktop / Claude Code / Cursor)        |        v   r3 MCP server (stdio)        |   +----+-----------------------+   |                            |embedded Redis            vectra local index(redis-memory-server,      (on-disk vector store auto-downloaded binary,    for semantic search, no external service)       no external service)   |   +--- optional ---> Mem0 cloud API (MEM0_API_KEY)                       cross-device sync, off by default
  • Embedded Redis — redis-memory-server downloads and manages a local Redis binary for you. It stores memory content and metadata. If it cannot start, r3 falls back to an in-process store.
  • vectra — a local, file-backed vector index used for semantic search. No network calls, no external vector database.
  • Mem0 (optional) — if MEM0_API_KEY is set, r3 also syncs to Mem0's cloud API so memories can follow you across machines. Without a key, nothing leaves your machine.

MCP client configuration

Claude Desktop

Edit claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/claude_desktop_config.json):

json
{  "mcpServers": {    "r3": {      "command": "npx",      "args": ["@n3wth/r3"]    }  }}

Claude Code

bash
claude mcp add r3 "npx @n3wth/r3"

Cursor

Add to .cursor/mcp.json in your project (or the global Cursor MCP config):

json
{  "mcpServers": {    "r3": {      "command": "npx",      "args": ["@n3wth/r3"]    }  }}

Restart the client after editing config. In a new conversation:

You: Remember that I prefer TypeScript and dark mode.AI: I'll remember that.
[new conversation]
You: What are my preferences?AI: You prefer TypeScript and dark mode.

Tools

ToolDescription
add_memoryStore content with optional metadata and priority
search_memoryQuery memories using semantic or keyword search
get_all_memoriesList stored memories with pagination
get_memoryRetrieve a specific memory by ID
update_memoryModify existing memory content or metadata
delete_memoryRemove a memory
deduplicate_memoriesFind and merge duplicate memories
cache_statsReport cache hit rate and storage stats
sync_statusReport Mem0 cloud sync status
optimize_cacheRun cache maintenance
import_memoriesBulk import memories

Enhanced mode (default, INTELLIGENCE_MODE=enhanced) adds:

ToolDescription
extract_entitiesExtract named entities from text
get_knowledge_graphReturn the entity/relationship graph
find_connectionsFind entities connected to a given entity

Configuration

Environment variables, all optional:

VariableDescriptionDefault
REDIS_URLUse an external Redis instead of the embedded oneembedded server
MEM0_API_KEYEnables Mem0 cloud syncunset (local only)
MEM0_USER_IDNamespace for memoriesdefault
INTELLIGENCE_MODEenhanced or basicenhanced

Example with cloud sync enabled:

json
{  "mcpServers": {    "r3": {      "command": "npx",      "args": ["@n3wth/r3"],      "env": {        "MEM0_API_KEY": "mem0_..."      }    }  }}

Comparison

r3mem0 (OSS)zep
Runs fully local with zero configyes (embedded Redis + vectra)requires a Postgres/vector DB you configurerequires a Postgres instance you configure
Needs an API key to try itnono (self-hosted) / yes (cloud)yes (cloud), or self-hosted setup
Optional cloud syncyes, via Mem0n/a (is the cloud option)yes

This table only reflects setup requirements observed in each project's own documentation, not benchmark performance or feature completeness. Verify against current upstream docs before relying on it.

Known issues

See LAUNCH_AUDIT.md for current limitations, including a native module build failure on some macOS setups.

Documentation

Full documentation at r3.n3wth.com.

License

MIT

Source: README.md at commit c4ef538

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v1.3.2LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts