Wink Pings

A single prompt turns a Kennedy speech into an archive-style video

BaoCut's new Skill can turn a speech or recording into an archive-style layout video from a single prompt: transcription, bilingual subtitles, historical images, and audio waveform visuals are all completed automatically. Some netizens commented that this is handing a requirements brief directly to AI.

Kennedy's 1962 speech at Rice University left behind that famous line: We choose to go to the Moon not because it is easy, but because it is hard.

Bao Yu posted a tweet showing this speech turned into an archive-style video. The speech footage occupies only part of the layout; beside it are a photo of the lunar surface, an audio waveform, and bilingual subtitles, with chapter titles made into subheadings. The whole thing looks like a historical documentary.

The key is the prompt. Bao Yu wrote only this passage, and BaoCut's Skill completed the transcription, translation, alignment, and layout:

Help me make a designed layout video, in the style of an archive or documentary: put the speech footage only in part of the frame, add bilingual subtitles, and if there are chapters, turn the chapters into subheadings; then search for and place some relevant historical images according to what he talks about. The whole thing should be beautiful, elegant, and cohesive. Horizontal format.

This workflow is not picky about material.

Audio works too. Sun Yat-sen's 1924 recorded speech "Principles for Saving the Nation" was laid out in the same format. On the left are the title and chapter text; on the right is the 1898 satirical cartoon "The Great Powers Dividing the China Cake"; below is the audio waveform, original audio 01:50 / 06:50.

Vertical format works too. The segment where Edison talks about the incandescent light bulb was made into a 1080×1920 vertical video. The subheading is "The lamp is only the beginning"; the chapter text reads "Think of the whole system first, then build the lamp"; at the bottom is a historical photo of Edison's light bulb.

The layout isn't fixed either. Bao Yu also showed another one: a 2003 speech by 20-year-old Dario Amodei at the Pasadena City Council opposing the Iraq War, made with a different color scheme and layout. He said he had AI choose the color combinations.

Some netizens commented that the most valuable part is actually that prompt. Archive style, showing the footage only in part of the layout, and matching historical images to what is being said—this is exactly the kind of brief that used to be written for a director.

Some short-video creators also said their biggest fear is timestamp drift and dropped words. BaoCut eliminates manual fine-tuning by aligning transcription with the timeline, which is what truly saves time.

BaoCut is an editing tool centered on editable video. Transcriptions, translations, subtitle layers, chapters, and voiceovers are all stored in the video, and users can continue editing in the editor at any time. Transcription means converting speech to text; aligned translation means matching the translation to the original audio sentence by sentence. The Skill file is public on GitHub and is used with the BaoCut App. It's not complicated: import your material into the App, invoke the Skill, and clearly describe the style you want.

URL: https://github.com/JimLiu/baocut/blob/main/skills/baocut/SKILL.md

What this kind of tool really changes is turning video making from operating a timeline into expressing intent. In the past, this kind of layout required a director, an editor, and a proofreader working together; now a single prompt is the requirements brief. AI does the tedious work, and the user only makes aesthetic judgments.

发布时间: 2026-10-26 00:45