AI & Media
Wondercraft
Built the MVP and the neural networks behind a YC- and ElevenLabs-backed AI audio studio.
The challenge
Wondercraft set out to be the Canva of audio: type, and get studio-quality podcasts, ads, and audiobooks in any language. That meant an MVP that could carry serious AI, agentic scripting, deterministic audio and video generation, and lip sync, without the roughness that kills an investor demo. It had to be real, and it had to be fast.
What we built
Agentic scripting flows
Agent-driven generation that turns a prompt or a document into a structured, editable script ready for production.
Deterministic audio and video generation
Generation flows tuned so the same input produces reliable, repeatable output: the difference between a toy and a tool.
Custom lip-sync neural networks
Bespoke neural networks for lip sync and several other core features, trained for quality the off-the-shelf options could not reach.
Full-stack platform
The end-to-end product around the AI: editor, media pipeline, and the platform that made it demo-ready and shippable.
ROI
Results
Outcome
The MVP did its job: it raised. Wondercraft closed a seed round backed by Y Combinator, ElevenLabs, Will Ventures, and Steven Bartlett, went on to serve tens of thousands of creators, and today powers some of the biggest names in podcasting. The neural networks we built are still part of what makes it work.
Technologies
