Ruby AI Early Access · v0.9 Get Early Access
Private beta · v0.9 · Waitlist open now

Your Mind, Amplified.
Your Data, Untouched.

Ruby AI runs powerful open intelligence directly on your Apple Silicon Mac. Zero monthly subscriptions. Zero cloud servers reading your life. Experience desktop-class reasoning, intelligent voice dictation, and seamless iPhone remote sync.

Which devices will you use?

One email when beta invites roll out. Nothing else — because we're a privacy company.

100% On-Device Inference Hardware Keychain Vault Zero Cloud Telemetry Apple Silicon Native

Section 01 · The architecture

The Brain & The Remote

Most AI apps are rented brains in someone else's building. Ruby inverts that: one hub does the thinking, every screen you own becomes a window into it.

The Mac Hub

Your Mac's unified memory executes the heavy load

Apple Silicon's unified memory executes open models from 8B to 70B parameters locally at lightning speed — zero token costs, no rented queue. Ruby keeps long documents, client context, and your whole workspace in local memory, ready in milliseconds.

M1 → M4 optimized 8B–70B local models $0 token cost

The Mobile Remote

Talk to Ruby from anywhere on your private gateway

Your iPhone or iPad speaks to your Hub over an encrypted personal gateway. Native voice dictation with zero audio lag, instant streamed responses, and your Hub's full context — from the kitchen to the boardroom, all without a cloud detour.

End-to-end encrypted Zero audio lag iPhone & iPad

Illustrative interface — current work in progress.

Ruby — Inbox Triage

A calm start on your desk

Ruby reads your local context, drafts the events, and executes — you approve in one pass. The model runs on your machine, so the whole loop happens in seconds, for free.

Streamed from token 1 — < 100 ms to first word, no server round-trip.
AppleScript bridge: Mail, Calendar, Finder, Notes, Shortcuts.

🎙️ Dictating — "move Thursday's three calls to next week…"

Say it walking down the hall

Hardware-routed dictation streams straight into Ruby — zero audio lag, full transcript formatting, and the same on-device brain answering from your Mac Hub over your private gateway.

The trip: iPhone → your Hub, on your network. Never a cloud.
Encrypted end-to-end and resolves only to your Mac.
Section 02 · The math

Why pay $20/month, indefinitely, to a chatbot?

Renting intelligence means the meter never stops — per seat, per message, per image. Ruby AI flips the model: one-time ownership with a transparent upgrade path when major new generations ship. You own it like you own your Mac.

Ruby AI
Typical cloud bot
Cost
✓ One-time ownership, transparent upgrades ($240+/yr saved)
✕ Eternal $240+/year subscription, per-seat
Privacy
✓ Data lives in local memory + your private iCloud database
✕ Conversations risk becoming training data on someone's server
Flexibility
✓ Local open weights free forever, optional BYOK for the extreme cases
✕ One provider, one bill, one rate card — no local fallback

🔑 Bring Your Own Key (optional, never required)

Need a heavyweight reasoning pass now and then? Plug in your own API key — your account, your data, your rates. Ruby stays local the other 99% of the time.

Google GeminiClaudeAny compatible endpoint
  • ✓One-time purchase. Pay once, use it forever. No metering — no per-message, per-token, or per-image trickle.
  • ✓Private by architecture. Your files and memories never leave local memory unless you send them to your own iCloud database.
  • ✓Zero cloud telemetry. No usage dashboards phoning home, no ad-network sync, no hidden analytics.
  • ✓BYOK for rare extremes. Gemini or Claude on your key, when you want it. Default mode stays 100% offline.
  • ✓Grows with your hardware. New Mac, new chips, bigger local models — free of a monthly ransom.
Section 03 · Everyday capabilities

An everyday assistant that never phones home

Everything below ships in the beta — running locally on your Apple devices with no API bill attached.

⚖️ AppleScript native

Executive Workflow

Triage unread email in one pass, draft calendar events, and fire system actions through native AppleScript automation. Ruby isn't a chat window bolted to your Mac — it's an operator with the keys to it.

tell application "Mail" → triage
tell application "Calendar" → draft event
Section 04 · Private beta

Reserve your place on the waitlist

Launching in about six months. Waitlist members get the closed beta first, founding pricing, and a referral link — every person you invite, you move up. A few details help us match beta builds to your hardware.

Get Early Access

Takes 30 seconds. Your details never leave our access-controlled server.

Protected by Cloudflare Turnstile. No spam — launch updates only.

  • ①Closed beta first. Waitlist members receive builds weeks before public launch, matched to your Mac hardware.
  • ②Founding pricing. Early list members lock the lowest one-time tier, forever.
  • ③Shape the product. Beta feedback goes straight to the build channel — features you ask for can ship in the same beta.
  • ④Referral priority. Your personal link moves you up for every confirmed signup, no limits.
Questions, answered plainly

Frequently asked questions

Does Ruby AI really work with no internet?

Yes. The assistant, memory, dictation, and creative tools run entirely on your Mac using Apple Silicon. Your prompts, files, and memories never leave the device — the internet is optional, not required.

How large are the local models?

Ruby's runtime executes curated open-weight models from roughly 8B to 70B parameters directly in unified memory, selecting the best fit for your chip and workload — refreshed as the open ecosystem improves, at no extra charge.

Can it use cloud models if I want them?

Optionally, via Bring Your Own Key. Connect your own Gemini or Claude API key for occasional heavy reasoning. It's always your key and your account — and the rest of Ruby stays fully offline.

Will I ever need a subscription?

No. Ruby AI is a one-time purchase with a transparent upgrade path for major new generations. No monthly fee, no per-seat charge, no token metering.

What hardware do I need?

Any Apple Silicon Mac (M1 or later) as your Ruby Hub, with the companion app on iPhone or iPad (iOS/iPadOS 17+). Intel Macs aren't the target — unified memory is the whole magic.

How does the iPhone talk to my Mac?

Over a secure, end-to-end encrypted personal gateway that resolves only to your Hub. Think of it as AirPlay for intelligence: instant, local, and yours.

Where is my waitlist signup stored?

In a private, access-controlled database on our own infrastructure — never sold, never synced to ad networks. We only use it to contact you about the beta.

🎉 You're in line

You're on the list.

Spot reserved for the closed beta line. Share your personal link — every signup it brings moves you up.