FishAudio | S2.1 Pro
Industry: AI Voice
Funding: Seed · $52M
Challenge & Solution
FishAudio builds open-source text-to-speech and voice cloning models, used by more than 8M people. In early July they made S2.1 Pro, their best voice model, free for every developer on the same API endpoint they already used.
A voice model is hard to launch. The product is sound, so the thing worth hearing isn't a thing worth watching, and a benchmark chart doesn't stop anyone scrolling. What does the launch video for a voice AI company even sound like?
So we had the product narrate its own launch.
FishAudio's own AI model voices the whole video. Anyone watching hears S2.1 Pro before they're told anything about it, which is a better demo than any chart.
The visuals followed from the voice: painterly textures, editorial type and a waveform system, with the benchmarks and stats animated in rather than pasted on screen.
We kept it to 45 seconds of pure motion with no shoot, cut for the feed first.
The launch video went out with S2.1 Pro on X and LinkedIn in early July, and a few weeks later FishAudio came back to us for their anniversary and funding announcement video.
Want this kind of eye on your own launch page?
Get a free teardown, back within 48 hours