PageCast
Stop reading it.
Watch it instead.
PageCast takes the page on your screen — including the ones behind a login — throws away everything that isn't the article, writes a script, and sends back a narrated video a couple of minutes later.
The video above was made from en.wikipedia.org/wiki/Hovercraft 6 slides1:10page to MP4 in 3 min 20 s
It works out what the page is actually about.
Every block and image goes to the model with its position, its class names and where its links point. Nothing is removed by a keyword rule, so a figure in the sidebar survives and a fundraising banner halfway down the article does not.
182
kept from the page in the video above
- 104 paragraphs, under all 32 headings
- 42 photographs, diagrams and drawings
- 28 picture captions
- The lead section, in full
67
removed, each with a reason
- 29
Main menu, search, contents, language linkssite navigation - 20
We owe you an explanation…donation banner and its buttons - 12
Site logo, tagline, duplicate meta imagepage furniture - 4
Categories, legal, footer navigationsite footer - 2
See also, related topicslink lists
Every removal keeps the model's own reason, listed beside you while you edit. Tick one back and the storyboard is rewritten to include it.
Nothing renders
until you've seen it.
The model drafts the slides and the narration. You get them as an editable storyboard, with a live preview of the finished frame beside every field.
- Twelve layouts
- Bullets, a pull quote, a big number, two columns, full-bleed photo. Change one and the preview redraws as you type.
- The page's own pictures
- Pick from everything the page had, at the largest size it offers. Nothing is generated, nothing is stock.
- Narration you can rewrite
- Each slide shows its word count and how long it takes to speak, so you can hit a length exactly.
It opens over
what you were reading.
No tab to hunt for, no download to find. The video plays on the page it came from, with subtitles, and the MP4 is yours to download whenever you want it.
Four pages, four videos, nothing touched up.
Each came straight out of the tool. They have sound.
What happens
when you hit render.
- Your browser does the reading
- It reads the page you are already signed in to. A server fetching the same address gets a login wall, a cookie prompt, or a different page altogether.
- We do the rest
- Filtering, script, voice, slides and encoding all run on our machines. Nothing to install, nothing to configure, and your laptop fan stays where it is.
- What travels, and what doesn’t
- The text and pictures from the page you asked about, and nothing else. No other tabs, no browsing history, no background collection.
- Honest about the state of it
- What exists today is a build that runs on one machine. The hosted version is what early access opens, and testers shape what it does first.
Tell us what you'd point it at.
We want a handful of people who read more than they have time for: researchers, engineers working through documentation, anyone who summarises the web for everyone else.