OS-level overlay for iPhone

See the AI
thinking.

Every time you talk to an AI agent, you're flying blind. Visible adds a real-time visual layer — surfacing what the agent is reasoning through, what's coming next, and what's happening on your camera.

Sales Agent — line open 2:34
thinking User asked about pricing. Checking knowledge base for plan details...
reasoning They mentioned "enterprise" — I should surface the Enterprise tier first, not the Starter plan.
next Outline pricing tiers → ask about team size → offer trial
camera ready — 2 objects detected
0 visibility you currently have
into AI agent reasoning

Voice AI is exploding.
Transparency isn't.

Retell, Synthflow, Vapi — voice AI agents handle millions of calls. Apple iOS 26 ships system call screening with AI. But the person on the other end of the call sees... nothing. Just audio. Invisible agents making decisions that affect you.

Visible flips this. It's the HUD for AI conversations — running at OS level on iPhone, surfacing the reasoning layer so you always know what's happening.

What's on screen

Live Agent Thinking

Watch the agent's chain-of-thought stream in real time. Not a transcript — the actual reasoning. Like watching DeepSeek R1 work, but during your call.

Voice Tree Navigator

See the conversation's next steps before they happen. Estimated wait times, where the call is going, what questions are coming. Navigate faster.

Camera + Multimodal

Snap a photo mid-call. The agent reasons over it in real time. Point your camera at a document, a product, a form — and have the AI work with what it sees.

Two-Way Agent UI

Don't just listen — edit. During agent-to-agent calls, you can intervene, correct, and redirect through the visual layer. You're not just a listener anymore.

How it works

01

OS-level integration

Like Granola, Visible runs at the iPhone OS level — capturing audio from any call, not just specific apps. Lock Screen activation. One tap to activate.

02

Stream reasoning in real time

The agent's chain-of-thought is captured and streamed to the overlay as it thinks. No transcript delay. No after-the-fact notes. Real time, all the time.

03

Multimodal layer

Camera input feeds into the same reasoning stream. The agent sees what you see. Snap, point, share — and the AI reasons with your visual context.

AI agents are making decisions
about your life right now.

You should be able to see what they're thinking.