Features

A full clipping pipeline,
in one Mac app.

Cutfarmer takes a long-form video — a podcast, a stream, a set — and turns it into finished vertical clips: found, framed, captioned, titled, and posted. Here is everything under the hood.

Clip discovery

Cutfarmer reads the whole episode — transcript, speakers, shot structure, audience reactions — and finds the moments that stand alone: a complete bit, a story with a payoff, a take that lands. Every clip opens on a hook and ends when the payoff finishes, cut to sentence boundaries, never mid-word.

Understands the episode

On-device transcription with word-level timestamps, speaker diarization, and shot detection give the engine a map of who said what, when, and on which camera.

Natural-length cuts

Clips run exactly as long as the moment needs — a 25-second punchline or a three-minute story — with a refinement pass that re-cuts both boundaries for a clean open and a complete ending.

Speaker-aware auto-framing

Turning 16:9 into 9:16 means deciding, frame by frame, who matters. Cutfarmer tracks every face, matches voices to the people on screen, and keeps the crop on whoever is actually speaking — through camera cuts, walk-ons, and crowded panels.

Captions and titles

Word-level captions

Bold, readable captions with word-by-word highlight timing, rendered into the video. Tap any word in the app to fix it; styles are template-driven so every clip matches your brand.

Hook titles

Every clip gets a short headline burned into the top band — written from what actually happens in the clip, positioned to never cover the speaker or the watermark.

Editing when you want it

Automation does the first pass; you stay in control. The three-phase workspace (Ingest → Edit → Export) gives you a per-clip timeline, framing inspector, caption editor, and a natural-language command bar — tell it "tighten the intro and keep the crowd shot" and it edits.

Publishing built in

Cutfarmer posts finished clips to the accounts you connect — TikTok and YouTube natively, with Instagram supported through the same pipeline. You sign in on each platform's own page; Cutfarmer never sees your password, and tokens live in a local file on your Mac.

Private by architecture

Everything — transcription, face tracking, rendering — runs locally on Apple Silicon. No cloud upload, no account system, no servers. The only thing that ever leaves your machine is the finished clip you chose to publish.

See how the pipeline works →