Leap

Creators were juggling a dozen tools to produce a single piece of content — one for audio, another for images, a third for video — each with its own account, billing, and export dance. The tooling was the bottleneck, not the ideas.

Talking to creators, the real cost wasn't any single generation — it was the context-switching between tools and the friction of paying for each. A single workflow that handled generation, payment, and storage would save more time than a marginally better model.

I built Leap as one platform for generating audio, images, and video, integrating multiple AI APIs behind a consistent interface. Bachs handles local and international payments, and a media storage layer keeps everything in one place.

Generative AI is slow and occasionally fails. I designed the workflow around asynchronous jobs with clear states and graceful retries, so a slow render felt like progress rather than a frozen screen.

Leap significantly reduced content production time for creators by collapsing a multi-tool workflow into a single platform — and taught me an enormous amount about shipping and running my own product end to end.

Building Leap solo — engineer, designer, and support desk at once — made me ruthless about scope. The product got better every time I cut something.