Script & scenes, written
Your topic becomes a structured script, then a scene-by-scene shot plan. Nothing to storyboard by hand.
Type a topic. It writes the script, builds the scenes, records a real neural voiceover, animates the captions, adds the motion — then exports the MP4 right in your browser. Vertical for Reels and TikTok, wide for YouTube. No render farm. No watermark.
How it works
There is no editor to learn and nothing to drag. You give it a brief; it hands back a finished file.
Step 01
One prompt box. Say what the video is about, roughly how long it should run, which language it should speak, whether it’s 9:16 or 16:9, and which of the three visual styles you want.
Step 02
Script, then scenes, then a real neural voiceover, then animated captions and background motion — assembled clip by clip and chapter by chapter. If a piece drops out, it retries and keeps going without you.
Step 03
The render runs on your own machine with WebCodecs — so there’s no queue, no render farm bill and nothing stamped across the corner. The file lands in your downloads, ready to post.
What’s inside
Not a pile of assets to assemble — the script, the voice, the visuals and the captions all come out of the same pass.
Your topic becomes a structured script, then a scene-by-scene shot plan. Nothing to storyboard by hand.
Microsoft edge-tts voices in many languages — a proper narrator, not a robotic beep-track.
Captions are timed to the voiceover and burned into the frame, so they survive every re-upload.
Chapter by chapter, clip by clip. Full explainers and documentaries, not just 30-second hooks.
Long builds run on their own and recover if something drops out mid-way. Close the tab, go do something else, come back and the video is further along than you left it.
Real stock footage when you want it to look shot. AI-generated images when you want it to look made. 2D animation when you want a cast of characters carrying the story.
WebCodecs renders the MP4 locally. Free every time, no watermark, no upload of your finished cut.
The same sign-in and the same shared credit balance as the rest of your Lumora apps.
2D animation
Pick the 2D animation style and your video gets actual characters. Every one of them is generated on a canvas, frame by frame. There is no stock artwork anywhere in it, which is why they can move, react and speak instead of sitting still.
bean
cat
ghost
robot
penguin
dragon
Output shapes
Choose the shape up front and the whole video is composed for it — framing, caption placement and pacing included.
Full-height framing with captions parked where thumbs aren’t. The format short-form platforms actually reward.
Room to breathe for explainers, documentaries and tutorials — the shape that suits a twenty-five-minute build.
Sample output
A full 70-second vertical cut. Nothing was edited afterwards — this is the file that came out of the export button.
Pricing
No per-export fee, no watermark to buy your way out of, no seat count. Credits come out of the same shared Lumora balance your other apps use.
Cost of a video
Exporting the MP4 is free — it renders on your own machine. Buy credits and manage billing at the shared Lumora account center.
| Video length | Credits |
|---|---|
| ~1 minute | 2 |
| 3 minutes | 6 |
| 5 minutes | 10 |
| 10 minutes | 20 |
| 25 minutes (max) | 50 |
FAQ
No. It runs entirely in a modern web browser at app.aivideogenerator.cc. You sign in with your shared Lumora account at id.lumorabuild.com — the same login you use for the other Lumora apps.
No. Export is free and runs locally in your browser using WebCodecs, so there is no render queue, no render farm and nothing stamped on top of your video.
Up to roughly 25 minutes. Long videos are assembled clip by clip and chapter by chapter rather than in one pass.
The build keeps going on its own and picks itself back up if a piece drops out. You can close the tab, come back later and carry on from where it got to.
Both. Choose 9:16 vertical for Reels, TikTok and Shorts, or 16:9 widescreen for YouTube. The framing, captions and pacing follow the shape you pick.
There are three visual styles: real stock footage, AI-generated images, or 2D animation. You pick the style and the generator builds every scene in it.
They are drawn by code on a canvas — no stock artwork. There are 28 characters, each with poses, walking, expressions and automatic lip-sync to the voiceover.
Real neural voices powered by Microsoft edge-tts, available in many languages.
2 credits to start a video, then 2 more credits per additional minute — so about 2 credits for a one-minute video, 10 for five minutes and 50 for twenty-five minutes. Exporting the MP4 costs nothing extra. Credits are your shared Lumora balance, managed at the account center.
Ready when you are
Describe it, walk away, come back to a finished MP4.