Soft Clipper Get Soft Clipper
Feature

It reads the whole video so you do not have to

Scrubbing a two-hour recording for the six good minutes is the part that kills the habit. Soft Clipper reads the transcript end to end and comes back with a shortlist.

📝

Transcript first

Existing captions are used when the source has them — that takes seconds. When there are none, the audio is transcribed with timestamps so every proposed clip maps to a real position in the video.

🎯

Clips, not chapters

A chapter tool splits a video into sections. This looks for something else: a segment that opens strongly, stands on its own, and stops before it drags.

🗣️

Ask for what you want

Type a request in plain English and get timestamps back. Useful when you know the moment exists and cannot remember where it was.

What a proposed clip contains

Every suggestion arrives as something you can judge in a second, not a mystery timestamp.

FieldWhy it is there
Start and endTrimmed to a sentence boundary so clips do not open mid-word.
Virality scoreA number for sorting the shortlist. Useful for ranking ten suggestions; not a prediction, in this tool or any other.
Hook titleA short punchy line you can drop straight on screen as the clip's headline.
Caption and hashtagsA ready-to-paste post caption in the language the video is in, so the clip is not sitting in a folder waiting for you to write something.
ReasonOne line on why this moment was picked, so you can disagree quickly.

You are never locked into the suggestion — every clip's in and out points can be nudged by hand before rendering, and the preview updates as you move them.

Frequently asked questions

How does Soft Clipper decide what is a good clip?
It works from a timestamped transcript rather than from the video pixels. The AI looks for a hook in the first seconds, a complete thought that does not need the surrounding hour to make sense, and a clean place to end — then returns the start and end times with a short reason.
Which AI model does it use?
Google Gemini, using your own API key. Gemini's free tier covers ordinary use; heavier use is billed by Google directly, never by us.
Can I tell it what to look for?
Yes. Prompt search takes a plain-English request — “the part about pricing objections”, “every time she tells a story” — and returns the matching timestamps instead of a generic best-of list.
What about videos with almost no talking?
There is a visual analysis mode that sends a low-resolution copy of the video to the model so it can watch the footage instead of reading words. It is slower, and it is meant for cooking, travel, gameplay and similar material.
Does my whole video get uploaded somewhere?
No. For a video with captions, only text is sent. When there are no captions, compressed audio is sent for transcription. Visual analysis is the one case where video is uploaded, and it is a low-resolution copy, only when you choose that mode.

Stop hand-cutting clips

Load a video, get back a set of vertical clips with captions. One payment, yours forever.

Buy for $49 — one time Windows 10 and 11 (64-bit) · 14-day refund
WhatsApp