All-In Clips

How it works

From a two-hour episode to a dozen short clips, tuned for each platform.

1. Ingest & transcribe. When a new episode is released, it is transcribed with word-level timing. Proper names are checked against a list of known spellings so captions say "Grok", not "Groc".
2. Find the moments. Candidate moments are ranked by how self-contained and shareable they are: a clear claim, a number, a disagreement, a payoff. Fragments that start mid-thought or need outside context are rejected.
3. Edit to vertical. Each moment is reframed to 9:16 with face tracking to keep the speaker centered. It gets burned-in captions, a title card, the speaker's name and loudness-normalized audio. Clips that come out wrong are automatically held back.
4. Write the copy. Titles, descriptions and the opening question in the comments are drafted from the transcript and checked so they only name people who actually appear in the clip.
5. Tailor per platform. The same moment is cut differently for each platform. Length, caption style, hashtags and posting time follow what has worked best on that platform.
6. Publish to our own accounts. Clips are uploaded on a schedule through each platform's official API, to accounts we own. On TikTok, the owner chooses each post, its privacy level and its settings from our dashboard.
7. Learn. We read the performance of our own posts (views, watch time, shares) to pick better moments next time.

The tools behind it

What we don't do

• We never post to anyone else's account. Every platform authorization covers one account that we own.
• We don't collect data about viewers. We only read statistics for our own posts.
• We don't edit what speakers said. Clips are cut at sentence boundaries and captions follow the audio.