← Funktionen

Turn a podcast or talk into captioned short clips

17. Nov. 2026 · 2 Min. Lesezeit · ViralShaker team

Tarife
Every plan. From 6 credits; each clip uses one content save, and a job that makes no clips is free.
Agent-Tools (MCP)
upload_media_from_urlestimate_clipsmake_clipsget_clip_jobshow_review

Give ViralShaker a long video you own, such as a podcast episode, an interview, a webinar or a conference talk, and it returns up to 8 vertical clips of the strongest moments. Each clip is reframed to 9:16, gets burned-in captions and an on-screen headline, and comes with a post caption, so it's ready to schedule.

What happens to your video

  1. Transcription. The speech is transcribed with timings, so captions line up word by word.
  2. Picking moments. AI reads the transcript and picks the parts that work on their own: a clear point, a strong opinion, a story with a payoff. Each clip gets a score from 0 to 100.
  3. Cutting. Each moment is cut from your source. Nothing is generated, so what people see is what was said.
  4. Reframing. The clip becomes vertical with one of three layouts: speaker (follows the face of whoever is talking), fit (the whole frame over a blur, for two-shots) or fill (a centre crop).
  5. Finishing. Captions are burned in, a headline goes on screen, and a post caption is written for each clip.

The clips land in your Library as ready content items. Review them, drop the ones you don't want, and schedule the rest to TikTok, Reels, Shorts, LinkedIn and the other platforms (see posting to every platform).

Limits and pricing

  • Source video: 1 minute to 1 hour.
  • Up to 8 clips per job, each 15 to 90 seconds.
  • Priced in credits from 6, depending on how many clips and how long. You see the most it can cost before it starts.
  • Each clip uses one content save. A job that finds no usable clips is free.

See pricing for the credits and saves each plan includes.

For AI agents

An agent imports the episode with upload_media_from_url (with temporary set, it's deleted a day later), gets the maximum price from estimate_clips, asks you, then calls make_clips with a max_credits cap. get_clip_job reports progress, and show_review puts the finished clips in front of you in the chat to approve and schedule. More on connecting an agent on the MCP server page.

Good to know

Only use videos you have the rights to. Clipping works best on speech: a solo episode or a two-person conversation gives better moments than a panel with crosstalk. For older short videos you already have, see repurposing old TikTok videos.

Häufige Fragen

How long can the source video be?

From 1 minute to 1 hour. It can be a podcast, interview, webinar, talk or any other video you own the rights to.

Does ViralShaker generate new footage for the clips?

No. Each clip is cut from your own video. AI only picks the moments, writes the headline and post caption, and transcribes the speech for captions.

What if my video has two people on screen?

Pick the layout that fits. Speaker follows the face of whoever is talking, fit shows the whole frame over a blurred background (good for two-shots), and fill crops the centre.

Do I have to keep the full episode in my Media Bank?

No. When an agent imports it with upload_media_from_url and temporary set, the episode is deleted a day later, after the clips are made.

Ähnliche Artikel