Captions are a commodity, and we'll quote them at cost. Description is the product — written by a producer against the actual slide, timed to the pauses, and recorded. Not a synthetic voice reading a machine draft.
Public universities in jurisdictions of 50,000+ must meet WCAG 2.1 AA on web and app content, video included.
Smaller public entities and special districts follow a year later, under the DOJ's April 2026 interim final rule.
Existing ADA obligations still apply today, and a back catalog takes months to clear, not weeks.
A lecture full of slides is the hardest content on campus to make compliant, and the one most programs underestimate.
Captions carry the audio. They do nothing for a chart, a formula, a diagram or the text an instructor points at and never reads aloud. For a student who can't see the screen, the most important minute of the lecture is silent.
WCAG 2.1 AA closes that gap with audio description, a second narration track written to fit the natural pauses and then recorded. That means scripting, timing, voice and mix. It's production work, not transcription — so the volume vendors solve it the only way a transcription desk can: generate the script from a machine read of the frame, hand it to a synthetic voice, and price it as the cheap line on the order form.
That version is available everywhere, and a student notices it inside thirty seconds. For us description is the core business. The same house that cuts broadcast work writes, times and records your track — and you choose the voice.
Anyone can generate a description. Almost nobody writes one.
| Criterion | Requirement | Transcription desk | OCM |
|---|---|---|---|
| 1.2.2 | Captions, prerecordedLevel A · accuracy, speakers, sound | Core | Included |
| 1.2.4 | Captions, liveLevel AA | Varies | On request |
| 1.2.5 | Audio description, prerecordedLevel AA · where the market splits | Machine script, synthetic voice | Producer-written, your choice of voice |
Title II web rule, 28 CFR part 35, subpart H, adopting WCAG 2.1 Level AA.
Transcription, speaker separation, visual indexing and first-draft description run on our own stack. Then an editor and a producer sign off, on every asset, no exceptions.
Send a folder or a Panopto, Kaltura or YouTube link. We inventory it, flag what needs description, and lock a schedule.
The pipeline drafts the transcript, separates speakers and indexes on-screen visuals in minutes.
Nothing ships from this stage.
Editors correct terminology, names and course vocabulary. Every asset.
AI drafts from the visual index; a producer writes and times the final pass, then records it, in a studio voice or a licensed AI voice for library-wide consistency.
Speaker IDs, non-speech audio, reading rate, sync, placement. A second-pass audit, then delivery back into your platform.
Per asset, from intake to a captioned, described file back in your platform, against a written SLA.
Choose by risk, then by voice. Captions are the easy half and everyone sells them cheaply — so we quote them at cost when they ship alongside description, and we compete where the work actually is.
The easy half, priced like it.
A description track written by a producer, not generated from a frame grab.
For the content a prospective student or a regulator actually watches.
Description is where the market splits, not where it stops. Every volume vendor will sell you a described track. What arrives is a script generated from the frame and read by a synthetic voice, priced as the cheap line on the order form — and at the top of the same rate cards, a human-voiced version costs several times what the machine one does. We start where they finish: the script is written against the actual slide, and the voice is a decision you make per library, not a surcharge you discover.
Priced per finished minute, by volume, at 100, 500 and 1,000 hours a year. Captions quoted at cost when they ship with description. Reduced archive rate for back-catalog rework. Rush available. A campus-wide framework once a standard is set.
Pick the department with the largest slide-heavy back catalog, where description actually matters. We caption and describe it at a fixed pilot rate against a written SLA. You get a measured cost per hour, a quality benchmark on the hard content, and a compliance record you can show counsel, before anyone commits to a campus-wide number.
Captioned and described. Cost per hour benchmarked on the hard content.
A single standard, priced by volume instead of negotiated department by department.
The same framework across your system, with your campus as the reference.
We've spent twenty years making video for networks, studios and the companies below. Accessibility is the same craft with a stricter spec.
Our own AI stack handles transcription, speaker separation, visual indexing and the first draft of the description. That's how a written, directed track costs what a generated one does elsewhere.
Description scripting, voice direction and mix happen in the same house that makes broadcast work. An editor and a producer sign off on every asset.
We've stood up in-house studios and trained client crews. If you want this in-house by 2028, we hand you the workflow.
The DOJ's April 2026 interim final rule moved large public entities to April 26, 2027 and smaller ones to April 26, 2028. That's more time to reach WCAG 2.1 AA, not a pause on the ADA obligations you already have.
A departmental back catalog takes months to inventory, caption and describe. Starting with a pilot now gives you a real cost per hour before budgets lock.
They're a good first draft, and they're our first draft too. On their own they mishear course vocabulary and names, don't identify speakers, and skip sound cues. They also do nothing for what's on screen. Nothing leaves our AI pass without human review.
A narrated track that describes important visual information the audio doesn't cover: slides, charts, equations, on-screen text. WCAG 2.1 AA requires it for prerecorded video (success criterion 1.2.5). Slide-heavy lectures almost always need it.
Most do now. Ask two questions about what you're buying: who wrote the script, and who is reading it. On the standard rate cards the inexpensive description line is a script generated from the video and voiced synthetically, and the human-voiced line sits several times higher — which tells you what the vendor thinks the difference is worth.
We're happy to be measured against whatever you have. Send us ten minutes you've already had described and we'll describe the same ten minutes. It's the fastest way to settle it.
Panopto, Kaltura, YouTube, or a shared folder. We deliver SRT, VTT, burned-in captions and a described audio track, back into your platform.
Yes. We've built in-house creative teams for clients before. If your goal is an internal accessibility team, we'll document the workflow, train your people and stay on as overflow.
Compliance dates as published by ADA.gov. This site is general information, not legal advice.
A rough number of hours and where it lives is enough. We'll come back with a pilot scope, a fixed price and a start date.