Showcase

Short video ad creative examples

Finished MP4s — footage, motion graphics, and a CTA card in one render. Not a raw clip.

Most text-to-video gives you a raw clip that still needs an editor. The videos below are finished ads: a planned hook-body-CTA arc, AI footage, animated headlines, brand-coloured accents, and a CTA card — delivered as one MP4 you can upload without a Premiere round-trip.

They work across LinkedIn feed, Instagram and Facebook Reels, TikTok, and YouTube Shorts. Landscape 16:9 and portrait 9:16 are both first-class outputs, not one cropped from the other. Browse the range, then adapt a brief.

Footage and motion graphics in one call

The pipeline plans a three-scene story from your brief, generates the footage, then composites motion graphics on top: animated titles, transitions, and a CTA card tied to your domain. The footage runs a few seconds; the overlay extends the final runtime. The file that lands is the ad, not a clip to finish.

Output is silent by default — feed ads autoplay muted — so add music or voice-over in the platform editor if you need branded audio. Voice-over at the API layer is on the roadmap.

Vertical-native for Reels and TikTok

The fastest-testing surfaces (Reels, TikTok, Shorts) reward variant volume. A 9:16 render composes hero subjects, headline animation, and CTA card for vertical specifically — cropping a 16:9 source would push the headline out of the safe zone. The MP4 is encoded for direct upload to every vertical placement without a re-encode.

FAQ

Is the output a finished ad or raw footage?

A finished ad. The MP4 includes footage, animated headlines, brand-coloured accents, and a CTA card — ready to upload. Runway/Sora-class tools give you a clip; this gives you the ad.

Which aspect ratios are supported?

16:9 (landscape) for LinkedIn and YouTube, and 9:16 (portrait) for Reels, TikTok, and Shorts. Both are first-class — not one cropped from the other.

Does it include audio?

Output is a silent track by default, which suits muted-autoplay feed ads. Add music or voice-over in the platform editor; API-layer voice-over is on the roadmap.

How long does a render take?

Typically 3–5 minutes end-to-end, dominated by GPU queue length. Off-peak it can drop under 3 minutes.

Generate your own

Open the public terminal with a starter prompt pre-filled, tweak it, and ship.