Three coordinated surfaces — a mobile assistant, a remote dispatch dashboard, and an autonomous business operator — sharing one memory, one model fleet, and one operator: you.
iOS · Android · Web · Self-hostable
Start with the pocket assistant, add remote dispatch when you need to run real work, graduate to CEO when you're ready to delegate a whole function.
The everyday assistant.
Chat, research, voice, and local inference on your phone. Works offline via Apple Foundation Models on iOS 26+ and llama.cpp on older devices.
Co-work with your machines.
Dispatch tasks to agents running on your servers, stream their terminals, and collaborate on long-running jobs from anywhere.
Autonomous business operator.
Multi-agent orchestration for real work — customer support, content, outreach, bookkeeping — with guardrails, approval gates, and full auditability.
Every Auto-Gnome tier is useful on its own. Together, they compound — the same memory, tokens, and operator identity move with you.
Plan, research, and voice-drive Auto-Gnome offline or online. Same account syncs across every surface.
Sign in to remote.auto-gnome.ai once, tap Pair on your phone, and the mobile app joins the same session — no passwords typed twice.
Fire off jobs to agents on your own machines. Watch progress stream live, attach a terminal, roll back if something goes sideways.
When a workflow matures, hand it off to a scoped agent team with budgets, approval gates, and full audit. You review, it executes.
Every bullet below is live code, tested, and in production — not a roadmap.
Apple Foundation Models
On-device Gemini-class inference on iPhone 17 Pro / iPad M-series, iOS 26+.
llama.cpp local runtime
CPU + optional Metal on older iPhones — no cloud round-trip needed.
OpenRouter integration
300+ cloud models behind one key. Free-text model field for bleeding-edge drops.
Gemini Live fallback
Streaming multimodal fallback with generous context.
BYOM vault
Bring your own API keys — per-provider, enable-gated, stored in device secure storage.
ElevenLabs streaming TTS
Low-latency voice with per-persona voice IDs (Athena / Hermes / Hades).
Local Whisper ASR
Offline speech-to-text for privacy-sensitive flows.
Unified voice relay
Server-side relay to Gemini / Anthropic / OpenAI without leaking keys to the client.
QR pairing
One-tap session handoff from the web to mobile, no passwords retyped.
mDNS LAN discovery
Self-hosted servers show up automatically on the configure screen.
Live terminal
Multi-client WebSocket PTY — attach and watch, or take the keyboard.
Dispatch Saga
Multi-step jobs with live progress events over SSE and compensating rollback.
Biometric gate
Face ID / Touch ID lock on the mobile app before remote access resumes.
End-to-end encryption
XChaCha20-Poly1305 + X25519 DH between paired devices.
Argon2id + JWT auth
Short-lived access tokens, rotating refresh tokens, TOTP + backup codes.
Session management
See every active session, revoke individually or all-at-once.
Full audit log
Every privileged action — pairing, dispatch, role change — recorded and queryable.
Self-hostable
Open Rust server binary. Run it on a Pi, a VPS, or your workstation.
On-device inference where possible. When cloud is needed, you bring your own keys — Auto-Gnome never proxies them.
The Rust dispatch server, the mobile app, and this marketing site ship with source. Run Auto-Gnome entirely off our infrastructure.
Every privileged action writes to the audit log. Biometric gate on mobile. TOTP + backup codes on the server. Zero silent actions.
Start with Remote in your browser — the Rust server, the Flutter Web dashboard, and the QR-pair
mobile flow are all live at remote.auto-gnome.ai.