Your POV,
Prompted.

What they see builds trust. What you see builds confidence.

Smart widgets layered on your outgoing video feed. A private AI-powered HUD (heads-up display) only you can see. On Zoom, Teams, FaceTime, Meet, and Webex.

macOS 14+ · Apple Silicon + Intel · Public Beta · 0.2.4

Desk Mode
2:34 PM CDT
☀️ 78°F
John Appleseed
Founder
Callout Q3 · NRR +38%
Control HUD · only you see this
What they see

Public widgets,
in the video feed.

Public widgets render directly into your outgoing video. Everyone on the call sees them. Designed to be tasteful and slow-updating — never distracting from the conversation.

Branding
Name card, role, company logo, social handle
Context
Local time, local weather
Identity
Pronouns, location, time zone
Data + utility
Live stock-quote ticker, QR code for sharing links
Callouts on the moment
Push a stat, headline, or emoji reaction onto your video with one click — three slots (lower third, side, top banner)
Reactions on your video tile
Six signals — ✋ hand raised, ❓ question, 👍 agree, 👎 disagree, 🎤 mic issue, ⏰ wrap up — composited on your own tile where the speaker can't miss them. ⌥1–⌥6 or a one-click popover.
Picture-in-picture (Pro)
Composite any window or screen region as an inset on your feed — show your work and your face at the same time
Background blur or replace (Pro)
Vision-powered person segmentation. Light, medium, or heavy blur — or drop in any image
What only you see

Private HUD,
on your screen.

Private HUD widgets render on translucent windows pinned to your screen. Invisible to other call participants. The HUD pops in when you join a meeting and hides when you leave.

AI Assistant
Choose your source: bundled on-device Qwen (no key needed) · BYOK Anthropic or OpenAI · or your company's Azure OpenAI, Bedrock, or OpenAI-compatible endpoint. Streaming · persistent history.
Pre-meeting briefing
Calendar-aware · attendees, history, invite notes, what's-next · AI-drafted (opt-in)
Action Items + Question Queue
Type follow-ups as they're said. With Pro+AI, the assistant adds items it overhears. One click pushes any item to a callout for everyone to see.
Live Captions + Room Vibe
Real-time transcription on a translucent HUD. Pro+AI adds Room Vibe — a sentiment read that warns you when the room is shifting and a "what did they just ask?" pill when you missed a question.
Calendar Peek
Next meeting at a glance with live countdown — no Cmd-Tab to the calendar app between calls
Teleprompter + Timers
Scrolling script under your camera. Agenda timer, hard stop, pomodoro — keep meetings on time.

Show your desktop.
Keep your face.

Switch your camera feed from your face to your desktop — or a single app window — without screen-sharing. Your video tile becomes your screen. Faster and more dynamic than Zoom/Teams share, and visible the same way on every call platform.

Desktop or a single window
Whole-display walkthrough, or pick one app window — the rest of your screen stays private even if it's behind the picked window.
Face-cam overlay
Optional circular preview of your camera in any corner. Three sizes. Audience keeps seeing you while you're demoing.
One click, mid-meeting
Floating button appears during calls. Click to toggle. Right-click for the window picker or to move the face-cam corner.
Desk Mode
Picture-in-picture
Pro

Show your work.
Stay on camera.

Composite a window or a screen region as an inset on your outgoing feed, with your face still primary. Where Desk Mode replaces your face with the screen, picture-in-picture keeps both — the audience never loses eye contact with you while you point at the thing you're talking about.

A window, or any region you draw
Pick a single app window, or drag out an arbitrary rectangle of your screen. Everything outside it stays private.
Sized and placed how you want it
15% to 30% of the frame, any of four corners, or drag it anywhere with a live editor that snaps to the corners. Size is a fraction of the frame, so it scales with your output resolution.
One key kills it
⌃⌥⌘P takes it off the feed instantly, from anywhere. And if the window you're sharing gets minimised, the inset freezes its last frame rather than showing whatever lands behind it.

Three steps. No Terminal.

1

Download & activate

Download the macOS app and activate the camera extension through System Settings. macOS asks for an admin password twice during setup — once to copy the app into Applications, once to approve the camera extension — the same as any Mac app that ships a system extension. Nothing after that.

2

Pick the virtual camera

Choose "promptpov Virtual Camera" in your meeting app. Works with Zoom, Microsoft Teams, FaceTime, Google Meet, Webex, QuickTime, and any other app that lets you choose a camera.

3

Configure widgets

Drag, snap, resize in the dashboard. Changes sync to your Mac in seconds. Your AI keys stay in your local Keychain — promptPOV never sees them.

Designed for people who work under IT policy.

No Terminal, no scripts, no MDM profile required. Choose where your data goes — and see it in plain English before you turn it on.

Standard install, then user level
Setup needs an admin password twice: copying the app into Applications, and approving the camera extension in System Settings. Both are macOS requirements for any app with a system extension, not ours. After that promptpov runs entirely at user level — no daemon, no sudo, no MDM profile — and every permission (camera, mic, calendar, screen recording) is granted by you through standard macOS prompts. If your Mac is managed and your account is Standard, you will need IT to install ittalk to us and we'll help.
On-device transcription
Transcription runs on your Mac. No audio bytes leave the machine. Live captions during the meeting and the post-meeting transcript both come from local inference — no cloud transcription service in the loop.
Keys stay on your Mac
AI provider keys live in your macOS Keychain. They never touch the promptpov backend. When you use them, AI calls go directly to your provider — no intermediary, and no changes to the billing or data-handling relationship you already have with them.
AI Governance Ready
Point promptpov at your Azure OpenAI deployment, AWS Bedrock / GCP Vertex proxy, internal LiteLLM or Portkey gateway, or any OpenAI-compatible URL. Works with whatever AI-governance layer your company already has in place — no code changes, no promptpov-side config.
Every AI call, logged locally
Local audit trail records every AI call — where it went, how big, how long, and whether it succeeded. Never the prompt or response itself. Export it for your IT reviewer any time.
Runtime enforcement, not just UI
Set the profile to local-only and the code refuses to send AI calls over the network — checked at every request, not just hidden behind a toggle. Verifiable, not just promised.
Sensitive titles auto-pause
Meetings whose titles contain HR, 1:1, legal, medical, interview, confidential, privileged, or attorney terms auto-pause transcription by default. Override in the Control HUD when you want it on.
Have questions from your IT team?
Send them our way. We'll walk through the data-flow log, the runtime guard, and the local-only mode with your security reviewer.
contact@promptpov.com →

Built for video meetings.

Priced for solo professionals.

Free
For trying it out.
$0 /mo
Public widgets (name card, time, weather, QR, social)
Basic HUD widgets (timer, teleprompter)
Desk Mode — desktop/window as your camera + face-cam overlay
· "Powered by promptPOV" watermark
· No AI features
Download free
Most popular
Pro
Everything except AI.
$19 /mo
or $190/year ($15.83/mo, save 17%)
All public widgets, no watermark
Full HUD (timers, teleprompter, calendar peek)
Desk Mode — desktop/window as your camera + face-cam overlay
On-device transcription (live captions + post-meeting transcript)
Pre-meeting briefing (attendees, history)
Email support
Subscribe
Pro + AI
With the AI copilot.
$29 /mo
or $290/year ($24.16/mo, save 17%)
Everything in Pro
AI Assistant — bundled on-device model, BYOK Anthropic/OpenAI, or your company's custom endpoint
AI-drafted pre-meeting briefing
Live AI: Room Vibe sentiment + "what did they just ask?" pill
AI post-meeting wrap-up + transcript polish
Priority support, early access
Subscribe
🆓 14-day Pro+AI trial — no card required.

Every new signup starts with 14 days of Pro+AI — the top tier, AI copilot included — automatically. No payment method at signup. When the trial ends, your account converts to Free — you'll never get a surprise charge. Upgrade from the dashboard whenever you're ready.

🎙️ Transcription runs on your Mac.

Meetings are transcribed locally on your Mac — no audio uploaded, no cloud transcription fees. Live captions during the meeting and the post-meeting transcript both come from on-device inference.

🔑 Bring a key, or don't.

Use bundled on-device Qwen models with no key at all — nothing leaves your Mac. Or bring your Anthropic / OpenAI key (stored in your macOS Keychain, never sent to promptPOV's servers, no inference markup). Or point at your company's Azure OpenAI, Bedrock proxy, or OpenAI-compatible endpoint.

👤 Per user, not per call.

Flat monthly subscription. No per-call, per-hour, or per-token billing. Use it on every meeting in your day or only the important ones — same price.

📅 Calendar-aware loadouts.

Tie a different widget set to each meeting type via title patterns. Sales calls surface the teleprompter; 1:1s pull in the action-items HUD; demos auto-enable picture-in-picture. Set up once, forget about it.

Ready to try it?

Public Beta · every new signup starts with 14 days of Pro+AI — no card required.