Shorts Layouts, Content Goals and Caption Styles in ClapClip: Every Setting Explained

The AI Shorts Generator in ClapClip for Windows has a small number of settings, and each one changes the result more than it looks. The content goal changes which moments are chosen; the layout changes which pixels survive; the caption controls change the one thing most viewers actually read. This guide goes through them in the order you meet them.

Key takeaways

• Three layouts, fixed sizes: vertical 1080 × 1920, square 1080 × 1080, landscape 1920 × 1080.
• Smart crop or fit. Crop follows the subject into the new frame; fit keeps the whole original frame with bars.
• Six content goals tell the analysis what counts as a good moment: viral, educational, sales, story, remix and music video.
• Captions are brand controls: font family, size, an accent colour for the highlighted word, and switches for the on-screen hook and emoji.
• Ask for up to twenty clips per run, or leave the count to the analysis.
• Everything renders on your PC; the analysis works from the transcript and a handful of small reference stills.

Layout and framing

Layout	Output size	Made for
Vertical	1080 × 1920 (9:16)	TikTok, YouTube Shorts, Instagram Reels
Square	1080 × 1080 (1:1)	Feeds that show square previews, LinkedIn, carousels
Landscape	1920 × 1080 (16:9)	YouTube, presentations, a clip that stays in the original aspect

Framing decides how a 16:9 source becomes one of these:

• Smart crop reframes around the subject so the speaker stays in the picture when the frame narrows. It is the default for talking-head sources.
• Fit keeps the entire original frame and pads the rest. Use it when the whole picture matters: a slide, a demo, a wide shot with two people, or a clip you intend to overlay your own background behind.

Rule: if the moment depends on anything at the edges of the original frame, choose fit. Smart crop is for faces.

Content goal

The content goal is a brief to the clip analysis. The same recording produces a different shortlist under each one:

Goal	What the analysis looks for
Viral	Complete, faithful moments with a clear audience payoff and a hook that stops the scroll
Educational	Clear claims, practical methods and conclusions a viewer can act on
Sales	Pain points, proof of value, objection handling and reasons to buy
Story	A person's situation, what changed and the real outcome, kept intact
Remix	Comedic lines, the setup that makes them land and the original reactions, without changing their meaning
Music video	Visual and sonic rhythm; multi-shot material that can be recombined to the beat

Pick the goal before generating, because it is cheaper than regenerating. A podcast about a product can be cut for education (how it works) or for sales (why to buy it), and the two shortlists barely overlap.

Rule: choose the goal by what you want the viewer to do afterwards, not by what the source is.

Clip count and language

Set how many clips to produce, up to twenty, or leave it at automatic and let the analysis return what it finds strong enough. Fewer clips with a higher bar beats a long list you have to prune.

The spoken language can be auto-detected or set explicitly; setting it avoids a misdetection on the first seconds of a file poisoning the whole transcript. Titles, hooks and publishing captions come back in the source language with Chinese and English versions alongside, so the studio can switch between them.

Captions

Short-form video is watched on mute, so the caption style is the brand. The controls are deliberately few:

• Font family and size. Default Arial at 56 on the caption canvas; the size is validated to stay between 16 and 120 so text never becomes unreadable or fills the frame.
• Accent colour. A single six-digit hex colour used for the word being spoken, so the highlight follows the voice. Set it to your brand colour once and every clip matches.
• Show hook. The on-screen title written for the clip, kept short enough to read on a phone. Turn it off for clips that open on a strong visual.
• Show emoji. Off by default; on for channels whose style uses them.

Captions are built from word-level timings produced by the local transcription, which is why the highlight lands on each word as it is said. They can be edited before export.

Rule: set font, size and accent colour once per brand and leave them. Consistency across clips is worth more than tuning each one.

Review, revise, export

Candidates arrive as a shortlist with their captions, hooks and hashtags. Watch each one muted first, then with sound. A clip can be revised (trimmed, its caption text corrected, its brand settings changed) and re-rendered without regenerating the whole set, and the result is saved so you can come back to it. Exports are rendered on your PC at the layout's full size.

The analysis itself works from the transcript and a handful of small reference stills so it can tell what is on screen when it scores a moment; the footage, the audio and the rendered clips stay on your machine. The clip-selection guide explains that pipeline step by step.

Bottom line

Three layouts, two framings, six goals, five caption controls and a clip count. The goal decides what gets picked, the layout and framing decide what survives the crop, and the caption controls decide what the viewer reads. Set the brand once, choose the goal per project, and spend your attention on the review pass.