AI video editing—inside your browser. Creators and AI agents work on the same real timeline. Voice, captions, vision effects, and export stay editable instead of becoming flattened output.
LOCAL INFERENCE with ONNX Runtime, WebGPU, and WASM · EDITABLE TIMELINE for real clips, captions, effects, and keyframes · PRIVACY BY DESIGN to avoid unnecessary uploads.
Launch Studio → · Explore the source →
🏆 Summer Bug Smash Winner — Clear the Lineup (2026)
One of five Clear the Lineup winners in DEV's Big Summer Bug Smash, powered by Sentry, for fixing a mobile click-through bug caused by unmounting a React portal on pointer down.
Official winner announcement → · Read the debugging write-up →
- Timeline Studio · Live demo — local-first browser AI video editing on a shared creator-and-agent timeline.
- Awesome GPT-6 Astra · Play games — a community-curated collection of GPT-6 Astra games, with playable demos and creator stories.
- Awesome AIGC Creative Contests — a continuously verified directory of active creative AI competitions.
- Vocal Remover Web · Live demo — browser-side vocal and instrumental separation.
- DeOldify ONNX Web · Live demo — old-photo colorization with ONNX Runtime Web.
- Depth Anything Web · Live demo — quantized monocular depth estimation in the browser.
I explore how far the browser can go as a serious AI application runtime. The interesting part begins after the model demo: graph optimization, GPU/WASM inference, workers, caching, media pipelines, temporal vision, responsive interfaces, and product decisions that make advanced tools understandable.
Right now I am focused on:
- browser-native AI video creation with privacy-friendly local inference;
- temporal computer vision that understands more than the first frame;
- reliable speech, caption, audio, and export pipelines;
- creative interfaces that keep AI output editable.