DashmadeDashCut

Native for Mac

Long video in. Short clips out.

Turn podcasts, interviews and talks into vertical clips ready to post. The speaker stays in frame, the captions land word by word, and your footage stays on your Mac.

Coming soon to the Mac App Store See how it works

How it works

From an hour of footage to clips worth posting

  1. 1

    Transcribe

    Word-level transcription with NVIDIA Parakeet, running entirely on your Mac.

  2. 2

    Label speakers

    Sortformer splits the transcript into speaker turns automatically. No thresholds to tune.

  3. 3

    Find moments

    A language model ranks self-contained 20–90 second spans on hook, impact, clarity, quotability and payoff.

  4. 4

    Review

    Preview each candidate as the real vertical crop with real captions. Nudge boundaries, approve or reject.

  5. 5

    Render

    Hardware-accelerated export with the speaker kept in frame and captions burned in.

Features

Everything between the long cut and the post

Face-tracked framing

A static crop when the speaker holds still, a smooth pan when they drift, a stacked split when two people are too far apart for one frame.

Animated captions

Word-by-word highlighting with five built-in presets and an editor for your own.

Every shape you post

9:16, 4:5 and 1:1 output, with a serial render queue and per-job progress.

Batch render

Queue every approved clip at once, skipping ones already rendered.

Transcripts and post copy

Export TXT, SRT, VTT or word-timing JSON. Generate social captions per clip.

What you see is what you export

Preview plays through the same compositor as the export, so there is no separate preview render to wait for.

Private by design

Your footage never leaves your Mac

Transcription and speaker labelling run on-device with models bundled in the app. There are no accounts and no uploads. Only transcript text goes to the language model you choose, and not at all if you run one locally.

Bring your own model

You pick the provider

  • Anthropic
  • OpenAI
  • OpenRouter
  • Groq
  • Ollama / LM Studio
  • Any OpenAI-compatible endpoint

Each provider keeps its own key in the macOS Keychain, and a one-click connection test checks the whole setup before you import anything.

Plans

DashCut Pro

One subscription unlocks everything that does work: import, transcription, speaker labelling, clip finding, rendering, export and social captions.

Monthly

Flexible, cancel any time.

Annual

Best value for regular use.

If a subscription lapses, your projects, transcripts, candidates and finished renders stay viewable. See current pricing on the Mac App Store.

Questions

What do I need to run it?

macOS 27 on a Mac. The speech models ship inside the app, so nothing is downloaded on first launch.

Do I need an API key?

For finding clips, yes: bring a key for the provider you prefer. Or point DashCut at a local model server such as Ollama or LM Studio and use no key at all. Transcription and speaker labelling never need one.

What leaves my Mac?

Your video and audio never do. Only transcript text is sent, directly to the provider you chose, when it finds clips or writes social captions. With a local model, nothing leaves at all.

What languages does it transcribe?

DashCut uses Parakeet TDT 0.6B v3, a multilingual model covering 25 European languages.

How do I cancel?

DashCut Pro is an auto-renewing subscription managed by Apple. Cancel any time in your App Store subscription settings. Your existing projects stay viewable afterwards.

Your next clip is already in there.

Coming soon to the Mac App Store