Everything PageToVid does

No timeline to assemble, no assets to source, no screen recorder to set up. The output is an MP4 you can publish, with a poster still and a subtitle file beside it.

Not stock, not a mockup. A browser loads your page at 2x and records it scrolling, clicking and hovering the things the script talks about — so the video shows the product actually working.

Scene animations for the moments a screenshot cannot carry: stat cards, before-and-after counters, comparisons, charts, pricing tables, quotes and calls to action, each with its own movement.

Generate photorealistic stills (Nano Banana) and short video clips (Veo) for beats no screenshot can show, and keep a recurring character — a person, a mascot, a product — consistent across every clip. A paid-plan feature.

Generated video scenes from Veo 3.1, Seedance, Wan or MiniMax, cut between real recordings of your site. One model per scene, priced in credits.

A generated presenter talks to camera in the corner while your site is recorded — and moves out of the way of what is being clicked.

Multilingual voices, chosen for character rather than language — firm, bright, upbeat, warm. Set the tone in words and the delivery follows.

Burned into the picture for feeds that autoplay muted, and emitted as a .vtt file for players that want a track. The timing comes from the recorded audio, not an estimate.

Accent colours are read from your site's own stylesheet, so the video matches the page it came from. Add a brand name for the opening and closing cards, or turn those off.

16:9 for the web and YouTube, 9:16 for Reels, Shorts and TikTok, 1:1 for feeds. Same source page, same script, framed for where it is going.

Rewrite narration, reorder or trim scenes, swap media, change the voice or the music, then re-render. Or leave a comment in plain language and have it applied for you.

Generate a set of videos from one site in a single pass — a landing-page promo, a vertical cut and a square cut — instead of starting three separate renders.

Point a show at an RSS, Atom or JSON feed and every new item becomes an episode — the same storyboard, fresh data — on a schedule you cap in credits per day. Nothing renders until you have previewed and confirmed the first run.

Twenty MCP tools let Claude, ChatGPT or your own agent inspect a page, plan a whole site, direct the camera at a specific element, correct a storyboard and re-render — with every claim's provenance and every scene's real captured page reported back.

Finished frames are graded automatically for unreadable text, overlaps and the wrong language, and the measured length is compared against the brief. A render that fails for a system reason refunds its credits.