Explainer Video Maker – Turn Articles into Videos

What Ribbi did

The user drops nothing but a link to an article breaking down the GPT-5.6 launch. Ribbi reads it and first offers three cut lengths — short, medium, long; with the ~2-minute medium cut picked, it distills the story into its load-bearing facts — the three-tier Sol/Terra/Luna lineup, ultra mode running 4 agents in parallel, Programmatic Tool Calling, and ChatGPT Work shipping the same day — then auto-classifies the material as facts-and-mechanisms content, which calls for a documentary tone. The ~400-word narration script lands in one pass: a cold open built on "check the benchmarks — this time they weren't bluffing," a body that explains Programmatic Tool Calling with an intern-who-phones-you-at-every-step analogy, and a closing thesis that the AI race is shifting from "who's smarter" to "who's cheaper, faster, and easier to slot into real workflows."

What Ribbi did

After generating 112.3 seconds of Chinese narration, Ribbi proofreads the subtitles line by line, fixing rendering glitches — digits split by spaces, "Programmatic" broken across lines, stray trailing text. It regroups the timeline into 23 semantically tight segments (total duration within 0.48s of target), then searches a royalty-free stock library clip by clip for matching visuals: OpenAI offices, data centers, close-ups of code, dashboards, city timelapses. When batch processing times out, it splits the work into three runs and switches mid-way to a steadier assembly method, padding the final segment down to a 0.01s error. The result: a 112-second documentary-style explainer with burned-in Chinese subtitles, every visual beat aligned to the narration.