Clip maker: Long-form to short-form (with & without speech)

I’d love to see a long-form to short-form video clipping feature added to Adobe Firefly.

The feature would allow users to upload a longer video and automatically generate multiple short-form clips optimised for platforms such as Instagram Reels, TikTok, YouTube Shorts and Facebook.

Ideally, there would be two clipping modes:

1. Speech-Based Clipping
Firefly analyses dialogue, captions, key phrases and topic changes to identify the strongest moments within a longer video. It could automatically create short clips with suggested hooks, captions and different duration options.

2. Visual / Non-Speech Clipping
This is particularly important for content that doesn’t contain dialogue, such as ASMR, beauty, product videos, craft content, tutorials, cooking, cinematic footage, behind-the-scenes videos and aesthetic content.

Instead of relying on speech, Firefly could analyse:

• Visual changes and scene transitions

• Camera movement

• Subject movement and actions

• Interesting or satisfying moments

• Changes in composition

• Audio peaks, sound effects and music beats

• Repetitive or low-interest sections that could be removed

• Natural start and end points for short-form clips

For example, I regularly work with ASMR content where there may be no dialogue at all. I’d like to be able to upload a 4–10 minute video and have Firefly identify several visually or sonically engaging moments and turn them into 15, 30 or 60-second clips automatically.

It would also be useful to have:

• Automatic 9:16 reframing and subject tracking

• Multiple clip suggestions from one video

• Adjustable clip length

• “More like this” regeneration for a selected clip

• Platform presets for Reels, Shorts and TikTok

• Automatic captions when speech is present

• Optional music/audio beat detection

• Hook or opening-frame suggestions

• Ability to manually refine the AI-selected in/out points

• Batch export of generated clips

This would make Firefly much more useful as an end-to-end AI video workflow, particularly for creators and brands producing large amounts of social content.

A visual-first clipping mode would also differentiate it from many existing AI clipping tools, which currently depend heavily on podcasts, interviews and talking-head content.