Turn Articles into Insight Videos

What Ribbi did

The user drops in an article link and one design-style reference image — not a single extra word. Ribbi follows the link and digests the article — Ilya Sutskever's first podcast interview since leaving OpenAI — then rewrites it as a thought-documentary narration script in a guided, conversational voice, slicing it into scene-by-scene segments, each with its narration copy and a visual design concept. Article comprehension, script rewrite, visual planning: all three staged before a single frame is generated — the whole film thought through before production starts.

What Ribbi did

Ribbi breaks the reference down into executable design language — solid royal-blue background, plump teal cartoon mascots, oversized rounded-type headlines filling the frame, a teal-on-blue color system — learning the design traits rather than copying the elements. The first five segments come out under this language, each with its own narration audio: five frames each telling their own part of the story, all looking drawn by one designer.

What Ribbi did

The locked style becomes a production template: segments 6 through 16 batch out under the same design language, each with its own narration audio. On the merge command, all 16 frames and tracks align into a 1920x1080 Full HD final delivery — the user says just two lines the whole way, and the rules set earlier govern every frame to the end.

What Ribbi did

Ribbi doesn't just nudge the speaking rate — it draws an emotion map across the content: surprise for cold opens, excitement for key declarations, calm for deep-thought passages, curiosity for questions, worry for warnings — then regenerates every narration line with its assigned emotional register and re-renders the final cut. Same frames, a voice with actual contour underneath, and the film suddenly feels human.