Paste a link. Get a video of what the page says.
Drop in a URL or attach a PDF. EggClip reads it before it writes a word, then narrates, builds every scene and times the captions. The video ends up being about the document, not its title.
150 credits on signup · no card · about five 30-second shorts
The source
example.com/2026/what-the-tide-gauges-actually-say
Read in full before a single line is written.
The figures in the piece become the figures on screen.
A PDF works the same way. Attach it instead of a link.
The video is about the document, not about its headline.
It reads the source before it writes.
The link or the file is part of the brief, so it gets fetched and read first. Then the script, the narrator, the scenes, the music and the captions are built from what the source said.
You bring
a URL, a PDF, or both
makes the other 5
one 9:16 file, posted now or from the calendar
You read the script before anything renders, and a failed render refunds itself.
Four renders from the same pipeline.
Real output from the demo workspace. A page or a PDF goes in at the top of this pipeline in place of a typed idea.
2 kinds of source: a link, or a PDF
64 narrators, previewable free
26 image styles, locked per project
2 channels it posts to
Four sources that work.
Anything with substance on the page. The reading happens before the writing.
- Paste the URL of a blog post Prompt to videoreads the page firstAI images
- Attach a PDF report or a whitepaper Prompt to videoreads scans too16:9
- Paste a documentation page and ask for an explainer Prompt to video60 secAI images
- Paste a news article and ask for the summary Prompt to videodated researchstock video
What comes out.
A video that cites the page it came from, with every file a channel needs.
Start free- Sources
- pasted links and attached PDFs, read before the script is written
- Scanned PDFs
- read as images as well as text, so a scan isn't a blank page
- Unreadable sources
- recorded as unavailable and stated, never quietly invented around
- Aspect
- 9:16, 16:9, 1:1 and 4:5
- Length
- ten seconds to three minutes vertical; chaptered long-form up to ten minutes
- Captions
- word-timed, seven preset families
- Voices
- 64 narrators, every one with a free preview
- Publish
- YouTube and Instagram, immediately or from the calendar
Questions people ask first.
Does it actually read the page, or just the headline?
It fetches the page and reads it before the script is written, and the same goes for an attached PDF. The script is built from what the source says.
Can it read a scanned PDF?
Yes. The document is read as images as well as text, so a scan or a multi-column layout comes through rather than arriving as an empty page.
What if the link is broken or blocked?
The source gets recorded as unavailable and the writer is told so, rather than the video confidently summarising a document nobody could read.
Can I combine a link with my own instructions?
Yes. The link or file is part of the brief, so you can paste a URL and say what you want made from it in the same box.
Will it copy the article word for word?
No. It writes a script for video from what the source says. If you want your own words spoken verbatim, paste them into the voice-over-to-video lane instead.
How long can the video be?
Ten seconds to three minutes vertical, or a chaptered 16:9 cut up to ten minutes for longer sources.
Other formats, same pipeline.
Turn the tab you have open into a video.
Paste the link and the first render comes out of the 150 credits that land at signup. No card.
Start free