Turn a Comic, Manga, Manhwa or Webtoon Into a Narrated Video
Comic to Video takes a comic — a PDF, a CBZ or ZIP, a folder of page images, or a long webtoon strip — and turns it into a narrated video: each panel held on screen over a blurred, slowly moving copy of itself while an AI narrator or cast tells the story. It is the same look the manhwa-recap channels use, built to get the crops and the reading order right where those rigs are sloppy: no bubbles sliced mid-word, no slivers of the neighbouring panel, and manga pages read right to left when you ask for it.
Pick a narration mode: a third-person recap narrator, a read-aloud cast that performs the dialogue voice by voice, or both together with the narrator bridging between performed lines. Write the narration in another language, choose a shape — wide 16:9, tall 9:16 for Shorts and TikTok, or square — and get back an MP4 with a subtitle track. Titles are yours to set: an opening card, an end card and a small on-screen credit are each a checkbox, and a video with none of them is a perfectly normal video.
What you make is private to your account and sits on a shelf on the page, ready to download. The Comics tab is the neighbouring tool for making a comic in the first place, and Document Reader and the Anime Dubber cover audio for documents and dubbing for video.
Open Comic to Video AI on Hard EA
What you can do
- PDF, CBZ/ZIP, a folder of page images, or a long webtoon strip in — a narrated video out
- Paste a link: a chapter, or a whole series on MangaDex, Webtoons or Tapas — pick the chapters and they are read in order, duplicate uploads merged
- Reading order detected for you: Japanese manga reads right to left, Western comics and webtoons left to right — panels and speech bubbles both
- Every character keeps one voice; covers, contents pages, credits and promo pages are cut so the story starts on its first page
- Panel-over-blurred-self camera move with correct crops and reading order, not sliced bubbles or panel slivers
- Three narration modes: recap narrator, read-aloud multi-voice cast, or both together
- Narration written in whatever language you pick, with a subtitle (.srt) file
- Optional title card, end card and on-screen credit — each one a checkbox, off or on as you like
- Wide 16:9, tall 9:16 for Shorts and TikTok, or square
- An API for AIs and scripts: the same tool behind an X-API-Key header
How it works
- Drop in a comic — a PDF, CBZ/ZIP, image folder, or webtoon strip — or paste a link to a chapter or a whole series.
- Pick a narration mode, a shape and a language, and say whether you want title cards.
- The pipeline reads every panel, writes narration sized to what is happening, casts the voices, times it and renders the video.
- Watch it on the page, then download the MP4 and its subtitles.
Frequently asked questions
What kind of comic can I put in?
A PDF, a CBZ or ZIP of page images, a folder's worth of pages, a single long webtoon strip, or a link: a chapter or a whole series on MangaDex, Webtoons or Tapas, a direct file link, or an archive.org item. A series link opens its chapter list so you pick which chapters to read; paid chapters (Webtoons Fast Pass, Tapas coins) are shown locked because the site will not hand them over.
Does it read manga right to left?
Yes. The reading order is detected from the pages — Japanese manga reads right to left, Western comics, manhwa and webtoons left to right — and it orders both the panels and the speech bubbles inside them. You can also set it yourself.
What are the narration modes?
A third-person recap narrator (the default, closest to the familiar YouTube format), a read-aloud cast that performs each character's dialogue in a distinct voice, or both together with the narrator bridging between performed lines.
Can it translate the comic?
The panels are read in their source language and the narration is written directly in the language you pick. Re-lettering the panels themselves — clearing and retypesetting the bubbles — is a later stage and will stand on its own as a separate tool.
Does it put a credit on my video?
Only if you ask it to. The opening card, the end card and the small on-screen credit strip are separate checkboxes fed by the title, creator and link fields; leave them blank and unticked and the video carries no titles at all.
How long does it take?
About two minutes plus half a minute per page — a 30-page chapter is around a quarter of an hour. The page shows how long is left and your place in line; you can leave while it runs, and the video lands on your shelf with an email when it is done.
Is output private?
Yes. What you make stays on your account; nothing is published anywhere by the tool itself.
Can an AI or a script use it?
Yes — the same tool is an API. With an X-API-Key header: POST /api/comic-video/detect-url to check a link, GET /api/comic-video/chapters/:source/:seriesId for a series' chapters, POST /api/comic-video with the files and links in order, then poll GET /api/comic-video/:id for the MP4, subtitles and poster. The full reference is section 19 of the developer documentation (llms-full.txt).