ClipViral AI

Problem

Turning a long video into shorts by hand means scrubbing the whole timeline for highlights, keyframing a 9:16 crop around whoever is speaking, syncing subtitles word by word and rendering each clip on your own machine.

What I built

The whole product, solo: the product design, the landing page, the web app and the processing pipeline behind it. Paste a YouTube link or upload an MP4 or MOV, and get back a batch of 1080×1920 clips with burned-in captions for TikTok, Reels and YouTube Shorts.

How it works

• Whisper speech analysis finds candidate moments in the dialogue; an optional prompt steers it toward punchlines, debates or lessons.
• Clips break at natural pauses (30–90 seconds), or you set an exact clip count.
• Face tracking keeps the active speaker centred in the 9:16 frame, switching between hosts and guests.
• Whisper word timestamps drive animated captions, with style presets.
• Clips render in parallel in the cloud; a browser editor allows trimming and caption fixes before export.

Key decisions

• One input on the hero: a YouTube link or a file, nothing else to set up first.
• The landing page shows the real workspace and a side-by-side source and output comparison.
• Credit-based pricing (5 credits per clip) across four monthly plans.
• Dedicated pages for each source and use case, from podcasts and streams to webinars and Zoom calls.

Where it stands

Live, with paid plans from $9 a month (100 credits) to $99 a month (2,500 credits). The Pro and Ultra plans include a 3-day free trial.