ADCraft
Human-quality audio description at scale: gap detection, timecoded scripting, hybrid voicing and embedded delivery, from the pipeline that described 50,000 videos in seven months.
Overview
Audio description from script to mixed track.
Audio description at volume is a scheduling problem as much as a creative one. Every video needs its gaps found, a script written to fit them, a voice recorded and a track mixed — and the whole thing has to be trackable when there are tens of thousands of videos and a fixed deadline.
ADCraft is the system behind that work. It detects candidate description gaps in the audio, gives describers a timecoded scripting view with the fit constraint visible, and carries the script through voicing, mixing and delivery with status per asset. It is the tooling that let one programme take more than 50,000 videos to WCAG conformance in seven months.
What it does, in order
Tabs pin the screenshot; text scrolls on mobileCandidate description windows detected from the audio track and offered to the writer.
Key features
Timed
Description gaps detected in the audio and scripts written to the exact gap, so nothing talks over dialogue.
Embedded
Standard and extended description tracks embedded in the player, not delivered as separate files nobody plays.
Compliant
WCAG and ADA media requirements met on every asset, with QC sampling by blind and low-vision reviewers.
Hybrid voicing
Professional narrators and AI text-to-speech combined for a consistent tone at catalogue volume.
Programme tracking
Per-asset status, ageing and throughput against the deadline, visible to the client.
Questions we get before the first call
See ADCraft on your own content.
Send a sample and we walk you through a real run, not a slide deck. Product inquiries land with the EdTech team as a tagged lead.