Publishing Accessible Videos
Publishing Accessible Videos
Do you want to publish a video, but you don't know how to comply with WCAG 2.1 AA accessibility requirements? This message is designed to help you understand your responsibilities so you can ensure the videos you publish are compliant.
In advance of the federal deadline of April 24, 2026, videos published on YouTube that do not meet compliance will be switched to "private" mode and will not be accessible to either ANR employees or the California public at large. (We can switch a noncompliant YouTube video to "public" once it is compliant and you'll keep all the old metrics.)
Any social media videos you publish from March 15 onward that are lacking compliant captions will be deleted because they can't be switched to "private." Unfortunately, any social media videos deleted this way means you'll lose all the engagement metrics associated with them.
There are two compliance deliverables required for every video that ANR publishes on any web platform. Some platforms have additional deliverables required; all of these are spelled out below. It is your responsibility to ensure you have all your deliverables ready and fully compliant before either publishing your video or requesting that Strategic Communications publish it for you.
YouTube videos now have 5 total deliverables required, 2 for compliance and 3 for YouTube.
- A finished video file
- A thumbnail image with proper brand identity, with a file size no larger than 1.5mb
- A title & description
- A properly formatted closed caption file or transcript stripped of timestamps [compliance deliverable]
- An alternate extended audio description (eAD) version of your finished video file [compliance deliverable]
For every social media video (TikTok, Instagram, LinkedIn, Facebook, X, BlueSky, etc.), there are now 2 deliverables required:
- A finished video file with compliant captions embedded [compliance deliverable] [1]
- A te[2] xt description with timecode stamps for each scene change that is inserted at the[3] head of the "comments" section [compliance deliverable]
Extended audio descriptions (eAD)
For visually impaired users, these are extended versions of videos which pause at each scene change so a voiceover can provide a WCAG-compliant description of the scene. The fastest and most economical means of obtaining an eAD requires a subscription at viddyscribe.com. This service uses generative AI to create an eAD file, which can then be downloaded. This process takes between 5 and 60 minutes, depending on the length and complexity of the source video. Viddyscribe will also simultaneously create a transcript of each description it has generated, with timestamps. This transcript can be downloaded and copy/pasted into the comments section of a social media video, thus meeting WCAG requirements. We recommend human review of all AI-generated transcripts.
Per UCOP, eAD for all non-YouTube social media videos must include a video description at the head of their comments section. A “video description” is a written description of each scene, with timestamps for when those changes occur. This allows visually-impaired users to access your videos with their screen readers. An example may be found at the following link.
Closed captions
These are written words that appear superimposed over the video to allow for the hearing impaired to read what spoken language occurs in a video. To meet WCAG 2.1 standards, captions need to be 99% accurate. Some social media applications (i.e., TikTok, Instagram) use AI to generate and embed captions into a social media video. Other systems such as Canva, Zoom, otter.ai, or rev.com provide generative AI services that can convert voice to text and output a transcript. Adobe Premiere Pro has a similar voice-to-text service built into the editor. Unfortunately, none of these AI tools can achieve more than 85% accuracy. This means they require human oversight and correction to be compliant.
Most of these tools require a subscription to access them, and free versions have even lower accuracy. You may opt to create them yourself (very time-consuming) or pay for professional transcripts through a service like rev.com. When formatted correctly, a transcript can be uploaded for a social media algorithm, such as YouTube's, to use in creating captions. Please note that Adobe Premiere Pro is particularly useful for both creating transcripts and embedding captions. You can also upload a transcript into Premiere if you're just embedding captions. Strategic Communications is developing training on this topic.
Here are your options for creating compliance deliverables:
- Cover your own cost of creating the compliance deliverables using tools like viddyscribe.com and rev.com (approximately $4 per minute of video in total)
- Pay approximately $4 per minute of video plus recharge to Strategic Communications for creating the compliance deliverables ($84.61 per hour)
Our vendors for these services are viddyscribe.com for eAD, and rev.com for human-generated transcripts. As of March 2026, compliance cost averages out to be approximately $4.00 per minute of finished video. However, we cannot simply pay by the minute; we must buy bundles of AI credits through Viddyscribe via a monthly or a yearly subscription. Furthermore, to have access to Rev's human transcriptionists, we must also pay them a monthly subscription. We encourage you to do your own cost research; however, whatever service you choose must guarantee an accuracy of 99% in captions.
If you have any questions, please feel free to reach out to Senior Videographer Ethan Ireland, Social Media Strategist Doralicia Garay, and/or Director Linda Forbes.
Downloadable PDF: