PowerPoint to video AI: turn a slide deck into a narrated video

To turn a PowerPoint into a video, export each slide as an image, copy the speaker notes into text, and send both to a video agent with a short brief. On Sume a run takes up to 30 images as attachments and the notes in input; a .pptx file is not an attachment type, so the export happens on your side.

Limits are from the Format API and Create a run docs, read 2026-09-29. Alibaba's Wan 3.0 API lists ppt and pptx among the files it can parse (API reference); Sume's wan-3.0 route does not take a document.

What do I export from the deck?

What if my deck has more than 30 slides?

Choose the 30 that carry the story, or split the deck into sections and make one video per section. Media URLs inside input count against the same budget of 30, so do not add slide links there as well.

Will the video keep my slide text exact?

Not by default. A model that redraws a slide can alter small type. If a slide must appear as designed, attach it as an image and say in the instruction that it is to be shown as is, then check the result frame by frame.

How do I add the voice?

Put the script in the notes and say in the instruction that the narration follows them. If you would rather record your own voice, say so and add it at the end. PDF to video AI shows the request body.