💡 Applications & Prompts

The 1920×1080 (16:9) format is the YouTube standard. The skill video-explicativo generates compositions in this aspect ratio with a palette #0D1321, Inter typography and smooth entrance animations, ready for direct upload.

YouTube is the world’s second-largest search engine — well-produced explainer videos generate long-term organic traffic. Mastering the 16:9 format is the foundation for any content strategy.

1920×1080 resolution, dark palette #0D1321, duration 2-5min, captions always on, CTA in the last third.

The 1080×1920 (9:16) format reverses the logic: everything needs to visually “fit” on a vertical smartphone. Larger text, faster cuts, a hook in the first 3 seconds.

Shorts, Reels, and TikTok have much more aggressive organic distribution algorithms than long-form YouTube—the same content, reformatted, can multiply your reach at no extra cost.

1080×1920 resolution, side safe zones, immediate visual hook, duration ≤60s (or up to 3min on Reels).

Retention studies on short-form platforms show a sharp drop starting at 90-120s. The skill uses ~110s as the default target for short videos — enough time to develop a point without losing the viewer.

High retention signals quality to the algorithm and increases organic distribution. Getting the duration right has a bigger impact than production quality.

Retention rate, watch time, LEAD=0.5 (opening silence), TAIL=0.9 (ending silence), cut pacing.

85% of videos in mobile feeds are watched without sound. Captions aren't optional—they're part of the composition. The skill generates the audio with pf_dora --speed 0.98 and the synchronized text appears as an overlay in the scene.

Captions increase retention, improve accessibility, and let the video work in environments without audio (office, public transit).

Text overlay, Inter font, minimum 4.5:1 contrast, synchronization with narration.

The video's last frame contains a call to action: URL, handle, product. The skill reserves the last 8-10 seconds for this "closing slide"—customizable via a variable in the prompt.

A video without a CTA is wasted traffic. Defining the CTA in the prompt avoids manual post-production and ensures visual consistency across the series.

Closing slide, destination URL, brand consistency, conversion.

The same narrated script can generate two distinct composition files: one with a 16:9 viewBox and another with 9:16. The narration is identical; only the scene layout changes. Half the work, twice the presence.

Content production is a bottleneck. Reusing the same script eliminates the need to write it twice and ensures a consistent message across platforms.

One script, two composition templates, DRY content, audio reuse.

An onboarding video explains a specific feature in ~90s — problem, solution, visual step-by-step. I replaced text documentation that nobody reads with an animated video users watch right in the product.

Onboarding videos reduce support tickets and increase feature activation. With the skill, you regenerate the video every time the feature changes — no post-production.

Feature scope, problem → solution, visual step-by-step, automatic regeneration.

Each course module can be a separate explainer video: an introduction to the concept, a demonstration, and a summary. The skill keeps the house style consistent across all episodes.

2–4 min micro-lessons have much higher completion rates than long lessons. Producing them in a series with the skill ensures visual consistency without requiring design work.

Single learning module, house style, cohesive series, FADE=0.45 in scene transitions.

Instead of a changelog markdown file no one reads, you generate a 90-120s video announcing what changed, why it changed, and how to use it. Shareable on Slack, email, and social media.

Video changelogs increase user engagement with new features and humanize product communications.

Visual release note, context + change + impact, multichannel distribution.

Architecture, technical decisions, a new library—any concept that would require a 30-minute meeting can become a 3-minute animated video the team can watch asynchronously.

Technical videos eliminate alignment meetings, remain as a permanent record, and help onboard new members months later.

Async-first, living documentation, team onboarding, animated diagram.

The video is generated from a text script. When the product changes, you edit the text and run the skill again—the video is regenerated from scratch in minutes, with no post-production.

Traditional recorded documentation quickly becomes outdated. The pattern "text as the source of truth + automatic generation" eliminates the maintenance cost.

Text as the source, automatic regeneration, script versioning, reproducible pipeline.

The default house style uses a palette #0D1321 + cyan + amber. But any color can replace them: just redefine the color variables in the prompt and the skill applies them throughout the composition.

Freelancers and agencies can offer customized videos for each client without creating separate pipelines — one skill, multiple brands.

Color variables, brand guidelines, white-label, custom palette.

The simplest prompt possible: topic, duration, and target audience. The skill handles the script, narration, composition, and export.

Understanding the minimum viable prompt lets you iterate quickly and use it as the basis for all subsequent variations.

Topic, target duration, audience, skill defaults, MP4 output.

A variation of the base prompt that instructs the skill to use the vertical composition template (1080×1920) with a duration ≤110s, a larger font, and an immediate hook in the first frame.

Specifying the vertical format in the prompt prevents rework—the skill automatically adapts the layout, timing, and pacing of cuts.

9:16 format, duration ≤110s, hook in the first 3s, text in a ≥48px font.

Replace the default house style with a custom palette. Specify the background, accent color, and voice tone — the skill reapplies them throughout the composition.

Ideal for adapting the video to a brand’s identity or creating themed series with different colors for each subject.

Hex background, accent color, narration tone, brand consistency.

Instructs the skill to display a specific CTA in the final frame — URL, handle, product, promotion. May include a QR code generated via composition or simple text.

Each video serves a different goal: capturing leads, selling a product, or gaining followers. The CTA needs to be intentional and specific.

End frame, short URL, conversion goal, urgency.

Controls LEAD (silence before narration), TAIL (final silence), and voice speed (--speed). Default values: LEAD=0.5, TAIL=0.9, speed=0.98. Increasing speed shortens it; decreasing it gives more breathing room.

Different platforms call for different pacing: Shorts needs speed, while technical documentation needs pauses to let information sink in. Adjusting these parameters is the difference between a video that engages viewers and one they abandon.

LEAD=0.5, TAIL=0.9, FADE=0.45, pf_dora --speed 0.98, cut pacing.

When only one scene in the video needs to change (wrong text, wrong color, bad timing), you instruct the skill to regenerate only that composition block — preserving the rest.

Regenerating the entire video to fix 5 seconds is wasteful. The surgical scene pattern saves time and keeps the rest intact.

Scene index, modular composition, non-destructive editing, selective rerendering.