AudioDesc
Automatic audio descriptions for video. Built for government channels, community media, and anyone who needs accessibility.
Requires an M-series Mac.
Loading latest version…The app is free. You bring your own Gemini API key for the AI analysis. A typical hour of video runs under a dollar.
How It Works
Drop a video into the app. It pulls still frames from the footage looking for moments where something visual appears that a blind or low-vision viewer couldn't know about from the audio alone. Slides, charts, graphics, title cards, that kind of thing.
When it finds one, AI reads what's on screen and writes a description. Then it finds a natural pause between sentences where nobody's talking and reads that description aloud in a synthetic voice that's intentionally distinct from the main audio so viewers know it's a description track.
Two Output Modes
Standard
Keeps the video the same length and slips descriptions into existing pauses.
Extended
Pauses the original audio, plays a longer detailed description, then resumes. Useful when there isn't enough breathing room to describe something complex.
You Review Everything
Before anything gets mixed into the final video, you review every description in a window that shows you a thumbnail of the frame, what the AI wrote, and how much pause time is available. Edit it, rewrite it, skip it entirely. Nothing goes in without your sign-off.
AudioDesc adds standard and extended audio descriptions to your video files. Your organization's specific accessibility and legal requirements are worth verifying with your own team or legal advisor.
Requires an M-series Mac.