Brand Hub / Social / Intro videos & sound

Intro videos & brand sound

Short introductory videos for social posts, in the three aspect ratios the platforms actually need — plus the single music bed used across everything.

These are introductory posts, not a marketing programme. They exist so the refreshed profiles have something to post and so we can see how the brand moves. The content programme — series, cadence, scripts — waits for the marketing/content team and the final website pass.
No voiceover, by design. Social video is overwhelmingly watched muted, so the message lives in on-screen text. It also means the narration-voice decision doesn't block posting. Card timings are already the segment boundaries a VO track would need, so narration can be layered on later without a rebuild.

Intro cut — 20.5s

Same six cards in every ratio. Copy is taken verbatim from the live messaging manifest, so the videos, banners and site never drift apart.
Landscape1920×1080
YouTube, X, LinkedIn and Facebook timeline posts. The default cut.
Square1080×1080
Feed posts on LinkedIn, Facebook and Instagram — takes more vertical space in-feed than landscape.
Vertical1080×1920
YouTube Shorts, Instagram Reels, Facebook Stories.

Card sequence

What appears, in order, and for how long.
#CardContentHold
1LockupCLOVITEK AI3.0s
2HeadlineWe build your product.3.0s
3HeadlineWe run your business.3.0s
4BodySoftware, marketing, and operations — handled by one AI-native team.4.0s
5BodyFrom idea to shipped, without hiring a department.4.0s
6EndCT+AI lockup · clovitek.ai3.5s

Brand sound

One bed, used across every video. Consistency is the point — do not vary it per post.
clovitek-brand-bed.mp3 — 45s, loudness-normalised to −19 LUFS so it sits under on-screen text without fighting it. Mixed at 55% with a 1.2s fade in and 2s fade out.
Narration voice — still open. The closest English voice available is Vitaly Pro, which carries a Ukrainian/Russian accent in English. That is not a blocker for developer and AI audiences, where specificity beats polish — but it is a brand decision, not a technical one, and nothing here depends on it. Whatever is chosen, use one voice and never change it mid-channel.
How these are produced. /root/scripts/gen_intro_videos.py — vector cards rendered by rsvg-convert and assembled with ffmpeg. All text is code, never model-generated (CT-5213), and the brand tokens and logo geometry are imported from gen_social_banners.py rather than re-derived, so the AI-pill and cropped-CT defects cannot reappear here. Regenerate with python3 gen_intro_videos.py.