What the AI actually does
“AI” gets stuck on everything these days, so here's precisely where it works in this maker:
- Neural voiceover. Kokoro and Piper text-to-speech models read your script in 49 languages with natural rhythm and intonation.
- Language detection. Paste a script in Spanish, Hindi or Arabic and a matching voice is selected automatically.
- Automatic scene building. Your script is split into scenes, with titles taken from each paragraph.
- Computer-vision sketching. Every image is analysed with edge detection and turned into pen strokes for the hand-drawn reveal.
- Speech-timed animation. Scene lengths, drawing speed, captions and music ducking are all calculated from the generated voice.
What it doesn't do: generate images or video footage from a prompt. You bring the pictures, which keeps your video original, on-brand and free of AI-art oddities.
Private AI, on your device
Most AI video tools send your script to a cloud server. Here, the AI models download to your browser once and run locally with WebAssembly or WebGPU. Your script never leaves your computer, there are no credits to buy, and it keeps working offline once the model is cached.
A fast AI workflow
- Draft your script — by hand or with your favourite AI writing assistant.
- Paste it into the Script tab and add your images.
- Press Animate my video, then Generate voiceover.
- Tweak any scene, pick a style, and export the MP4.
Try it in the animated video maker, and see all AI voices.