God's Eye View – Open-Source Live 3D Earth Tracker
Voice Control in God's Eye View: Talk to the Globe

Published 2026-10-12

Voice Control in God's Eye View: Talk to the Globe

The hands-free voice agent: how it works, OpenAI Realtime API setup, what it costs, and tips for good results.

One of God's Eye View's signature features is the hands-free voice agent: you talk, the globe moves. Powered by the OpenAI Realtime API, it turns the command center into something out of a sci-fi film. Here's how it works and what it costs.

What the voice agent does

With an OpenAI API key added in the app's settings, a voice mode appears. You speak naturally — "take me to San Francisco airport," "show me earthquakes from today," "follow that flight" — and the realtime agent interprets the command, moves the camera, toggles layers, and responds by voice. It understands the app's full context: layers, tracked objects, camera position. The demo moment everyone shares is asking it to find the nearest airborne flight and riding along without touching the mouse.

Setup

You need an OpenAI API key with access to the Realtime API. Paste it into the app's settings panel; the key routes through the app's hardened server-side proxy (SSRF protection, per-provider budgets) rather than living in the browser. Once saved, the voice interface activates. A decent microphone matters more than anything else — the agent is only as good as what it hears.

What it costs

The Realtime API is usage-billed by OpenAI — you pay per minute of voice conversation at their standard realtime rates. Casual exploring costs little; leaving it chatting for an hour adds up. The proxy's per-provider budgets help you cap spend. Everything else in the app remains free; voice is the one feature with a real meter running. Check OpenAI's current realtime pricing before a long session.

Tips for good results

  • Be specific about places: "SFO" works, "that airport near the bay" doesn't.
  • One command at a time: let the camera finish moving before the next instruction.
  • Name layers directly: "turn on the satellite layer" beats "show me space stuff."
  • Quiet room: background noise confuses any voice agent; this one's no exception.

Is it worth it?

If the cinematic factor is why you're here — absolutely. Typing coordinates is fine, but telling the globe where to go is the closest the project gets to its sci-fi promise. If you're budget-conscious, everything else works keyless, and the layer guide shows how much the silent version already does.