ViddyScribe

For the 285 million people worldwide with visual impairments, accessing video content can be challenging. Our goal was to create a solution that generates professional-grade audio descriptions automatically.

We researched existing audio description solutions and identified key pain points in current workflows, focusing on quality, speed, and cost-effectiveness.

We leveraged Google's Gemini API for scene understanding and natural language generation, combined with advanced audio processing for seamless integration.

We developed a streamlined platform that allows users to upload, process, review, and share their videos with professional audio descriptions in minutes.

The platform serves whoever owns the video. A blind viewer can't make a YouTube creator add audio description, so we built Realtime for the viewer instead. Paste a link or search for a video and start listening right away. Gemini watches the scene and describes it in the natural pauses between dialogue, so it never talks over anyone.

The industry calls them standard and extended audio description. Each option says what actually happens, whether the video pauses, and keeps the industry term in brackets for people who already know it.

Try Realtime ↗

For more than 6,000 people and organizations in over 130 countries.

Adopted across universities and government teams in the US to make video content accessible at scale.

Recognized as the Best Web App for innovative use of Gemini API in accessibility solutions.

Highlighted by the Chrome Developer team for creating an elegant solution to a real user problem, with particular praise for our user-first approach and consultation with visually impaired users.

Read the Chrome Developers post ↗