Grok Bot
Add it as a Cursor-format plugin on a shared bot computer. No API key, and it runs locally.
NEWThe relaunch: one memory for every agent, bot and chat window
Rent the LLM.
Rent the LLM.
Own the memory.
Now one memory for every agent, bot and chat window.
Local-first and governed. It says “I don’t have that” instead of guessing.
npm install -g superlocalmemory then slm setup · Windows, Linux and macOS
# Real output from a fresh Mode A install, trimmed
$ slm remember "Staging DB is Postgres 16 on port 5433." $ slm remember "Staging DB moved to Postgres 17 on port 5434." --replaces 0f57a17af3dd46e5 Replaced ✓ 1 fact(s) of 0f57a17af3dd46e5. $ slm recall "which port does staging postgres use" 1. [0.68] Staging DB moved to Postgres 17 on port 5434.
The old memory is kept and the change is undoable.
Works with the tools and bots you already use
Claude Code
Codex
Cursor
VS Code
Copilot
Antigravity
Windsurf
Zed
JetBrains
Gemini CLI
Hermes
ChatGPT
Claude
ComposioGrok Bot
Muse Bots, web apps and agents NEW Every agent. Every bot.
Every agent. Every bot.
Every chat window.
The same governed memory now reaches your bots and chat apps. Your laptop keeps one outbound link; the database never leaves it.
ChatGPT
Connect through Web access: sign in with GitHub, then choose read-only or read and save.
Muse
Connect through Web access and get the same governed recall a local agent gets.
Claude on the web
Claude Code on the web
Composio
MCP
Any remote MCP client
- 1Turn on Web accessAdd an app under Connected apps.
- 2Sign in onceGitHub OAuth. Read or save, per app.
- 3Calls reach your laptopThrough a gateway, on the same governed paths.
Your laptop must be online. Free daily allowance per connection. SLM works fully without Web access. See every host and plugin →
Why SuperLocalMemory Memory you can govern, trust
Memory you can govern, trust
and check before it speaks.
01 · Governed memory
Rules, not just similarity
Roles, scopes, GDPR erasure with receipts, retention rules and a hash-chained audit log.
See the governance → 02 · Recall you can trust
Memory that cannot quietly get worse
Five channels, a learned ranker with shadow tests and rollback, and time travel.
See how recall works → 03 · Answer check
It says “I don’t have that”
Laya on your Mac or Jev online judges whether the memories actually answer.
See Laya and Jev → Answer check NEW A judge for your memory.
A judge for your memory.
Laya on your Mac. Jev online.
Recall ranks by relevance. The answer check decides whether the top memories actually answer the question.
Laya: private, on your Mac
A small on-device model, about 1.1 GB. Nothing leaves your machine.
- Apple Silicon Macs
- Set up from the dashboard
- No key, no internet after download
“When is Project Kestrel going live?”
✓The memories answer it
Jev: online, on any desktop OS
Your own TypeSafe or OpenRouter key and an explicit consent. By default it sends the question and the top 3 memories.
- Off until you consent
- Same verdict as Laya
- One judge runs at a time
“How much did the Kestrel pilot cost?”
✗I don’t have that
Jev ranking: the answer comes first
An optional switch reorders the top results. Nothing is removed or hidden.
62%answer first, ranking on
49%answer first, ranking off
+0.4 sper recall
Also in the box Bounded loops, nine adapters,
Bounded loops, nine adapters,
multi-agent memory, governance.
01
Bounded loops
Iteration ends when an independent gate passes, never when the agent says it is done.
slm loop · /slm-loop · MCP tools02
Nine framework adapters
LangGraph, LangChain, LlamaIndex, CrewAI, AutoGen and four more share one local store.
See the adapters →03
Multi-agent memory
See which agent wrote each memory, and every bounded-loop run, in the dashboard.
04
Governance & EU compliance
Roles, sign-in, GDPR export and verified erasure, and an EU AI Act posture report.
See the controls → Full package
Eighteen capabilities. One local runtime.
Profiles and scopes
Isolated workspaces with personal, shared and global memory.
Knowledge graph
Entities, relations and graph-informed recall.
Learning and ranking
A ranker that learns from outcomes, with a shadow test and rollback.
Skill evolution
Measured skill revisions with lineage, under a budget.
Cache and compression
Exact cache and reversible compression, no proxy needed.
SLM-Mesh
Messages, locks and shared state across agents.
MCP, CLI, hooks, skills
One engine behind every surface.
Operator console
Dashboard for memories, graph, health, mesh and settings.
Opt-in ingestion
Gmail, Calendar and transcripts, off until you connect them.
Laya and Jev
A judge that says “I don’t have that”.
Nine memory kinds
Decisions, rules and how-tos. Standing rules load each session.
Bots and web apps
ChatGPT, Claude on the web, Muse and Composio.
Time travel
Ask what was true, and what was known, on any date.
Saved views
Reusable questions and summaries that cite their memories.
Reviewed corrections
Replace, review and roll back, with restore points.
Switch models, repair
Change embedding models and repair the store, with undo.
Backup and remote
Encrypted cloud backup, and memory served over TLS.
Plugins for every host
Fifteen skills, four agents, twelve IDEs, one rulebook.
Published LoCoMo evidence arXiv:2603.14588 ↗
Mode A · Retrieval
74.8%
10 conversations · 1,276 questions · local retrieval with GPT-4.1-mini answer synthesis
PUBLISHED V3 PROTOCOL Mode A · Raw
60.4%
10 conversations · 1,276 questions · local embeddings, retrieval, and zero-LLM answer construction
PUBLISHED V3 PROTOCOL Mode C
87.7%
Conv-30 · 81 questions · cloud embeddings plus GPT-4.1-mini answer generation and judge
PUBLISHED V3 PROTOCOL Published paper evidence (arXiv:2603.14588), not a fresh run of the current package. Mode B has no separate published run.
Quick Start Two commands to install.
Two commands to install.
One to connect.
# Install (Node 18+, Python 3.12+) — or: pipx install superlocalmemory $ npm install -g superlocalmemory
# Setup — pick Mode A to keep everything on this machine $ slm setup
# Connect your editor — claude-code, codex, cursor, windsurf, zed ... 12 IDEs $ slm connect cursor
Runs on Windows, Linux and macOS. No Docker, no required graph database, no API key.
# Setup — pick Mode A to keep everything on this machine $ slm setup
# Connect your editor — claude-code, codex, cursor, windsurf, zed ... 12 IDEs $ slm connect cursor
Runs on Windows, Linux and macOS. No Docker, no required graph database, no API key.
AGPL v3 · Free local core · Web access has a free daily allowance and is optional
Mathematical Foundation
Geometry, not guesswork.
Four public arXiv preprints, not venue-reviewed.
01 · Retrieval metric
dFR(p, q) = arccos( Σ √(pi · qi) )
Fisher-Rao Distance
Dense candidates use cosine similarity; Fisher-derived terms refine later scoring.
arXiv:2603.14588 →02 · Lifecycle model
∇γ ∂tγ = 0 · Expp(v)
Riemannian Lifecycle
Lifecycle combines policy, observed use and decay.
arXiv:2604.04514 →03 · Compression theory
H(X|Y) ≤ H(X) · I(X;Y) ≥ 0
Information-Theoretic Compression
Safe mode preserves JSON and code; reversible storage is verified separately.
Product evidence boundary → Integrations NEW
Every host, one memory.
Plugins for Claude Code, Codex, Antigravity, VS Code / Copilot and Hermes; a Grok Bot plugin; and slm connect for 12 IDEs.
Framework memory adapters
Nine adapters wire each framework’s native memory interface to your local SLM store.
| Framework | Package | What it wires |
|---|---|---|
| LangGraph | langgraph-superlocalmemory | BaseStore long-term memory |
| Semantic Kernel | semantic-kernel-superlocalmemory | VectorStore / RecordStore |
| Microsoft Agent Framework | agent-framework-superlocalmemory | Memory + conversation history providers |
| LangChain | langchain-superlocalmemory | BaseChatMessageHistory + retriever |
| LlamaIndex | llamaindex-superlocalmemory | Chat store |
| CrewAI | crewai-superlocalmemory | ExternalMemory storage backend |
| AutoGen | autogen-superlocalmemory | Memory provider |
| Google ADK | google-adk-superlocalmemory | Memory provider |
| OpenAI Agents | openai-agents-superlocalmemory | Memory provider |
Claude CodePlugin
CodexPlugin
AntigravityPlugin
VS CodePlugin
GitHub CopilotPlugin
HermesPlugin
Grok BotCursor-format plugin
Cursorslm connect cursor
Windsurfslm connect
Zedslm connect
JetBrainsslm connect
Gemini CLIslm connect
Claude Desktopslm connect
OC
OpenCodeslm connect
Cont
Continue.devslm connect
Also in use
MCP Protocol ChatGPT Desktop Perplexity Continue.dev GitHub Actions Docker n8n Goose OpenClaw nanobot Shell scripts
Common Questions
Frequently Asked Questions
Start now
Own the memory.
Two commands. Free, open source, local-first.
