AI chat that works while traveling — offline on flights, abroad, anywhere
The short answer
Yes — you can chat with AI in airplane mode. Download a model to your device once (from about 600MB), and ShadowTalk AI answers on your own GPU with zero network: no roaming, no SIM, no hotel wifi needed. Anything you send while offline queues in durable local storage and syncs automatically when you land. No app store, no account, no card.
Travel is where cloud AI fails hardest: airplane wifi that barely opens a webpage, roaming charges, hotel captive portals, dead zones on trains and in the mountains. This guide covers what actually works mid-flight, what to set up before you leave, and the honest limits — because offline-AI claims deserve fine print too.
The travel problem: AI dies exactly when you need it
Cloud assistants need constant connectivity, and travel strips it away: inflight wifi is slow and expensive, roaming data is costly and metered, hotel networks hide behind captive portals that block apps, and trains, mountains and rural routes have dead zones by design. The fix is local compute — a model that lives on your device and answers without asking any server.
That's what ShadowTalk's on-device mode is: WebGPU inference in a browser tab, running on your laptop or Android phone with zero network egress. The app installs as a PWA in one tap — no app store — and the whole shell works offline once installed.
What works mid-flight (and at 30,000 feet)
With the model downloaded before boarding:
- On-device chat is unlimited and network-free — drafting, translating, summarizing, planning, Q&A at the speed of your GPU
- Messages you send while offline queue in durable local storage and sync automatically when you land — end-to-end encrypted when Private Mode is on
- Your full local chat history stays readable offline — review notes and drafts from the ground
- The installed PWA reloads offline — no first-page stall, no app-store dependency
Prep checklist before you leave
Five minutes on good wifi covers the whole trip:
- Install the PWA and download the model tier your device suggests (600MB-2.6GB; the manager pre-checks storage quota)
- Use Chrome or Edge 113+ on laptop or Android — the app pre-checks WebGPU support for you
- Add provider API keys if you want frontier models the moment you reconnect (BYOK, at cost, no markup)
- GPU work draws battery — plugged-in seats help; a nano tier runs comfortably on modest hardware
Honest limits on the road
What offline AI can't do, stated plainly:
- No live web mid-flight — flight statuses, news and search need a connection; do research before boarding
- Frontier cloud models need the network — the on-device tier trades capability for independence
- iOS: the PWA installs on iPhones, but WebGPU support varies by iOS version — on-device mode targets Chrome/Edge on desktop and Android first
- The very first install needs a real connection — captive portals can block it, so set up before travel, not at the gate
Set up travel-ready offline AI before you fly
- 1On good wifi, open shadowtalk-ai.com in Chrome or Edge 113+ and install the PWA from the address bar.
- 2Open the model manager and download the suggested model tier once — 600MB-2.6GB depending on your device.
- 3Verify airplane-readiness: turn on airplane mode and confirm chat still answers (it runs on your GPU).
- 4Optionally add your provider API keys for instant frontier-model access when you reconnect.
- 5Fly: chat offline as much as you like — queued messages sync automatically on landing, E2EE with Private Mode on.
Frequently asked questions
Keep reading
Free to use — no account, no card.
Works offline. Syncs when you're back. Your keys, your models, your chats.
Install before you fly — free