Captions and transcripts for accessibility (ADA, WCAG, EAA): 2026 guide
Video captions and transcripts are required by ADA, WCAG 2.2, and EAA — and ignored by most teams until a lawsuit. Practical 2026 guide to which rules apply, who they apply to, and how to ship compliant captions fast.
One in five people worldwide has some form of hearing loss. Add in the millions who watch video without sound on public transit, in open offices, or while scrolling social media, and the audience for captions grows even larger. Yet a significant portion of online video still ships without accurate captions or transcripts.
This is not just an oversight. It is a legal and business risk that is becoming harder to ignore. Accessibility regulations are tightening globally, platforms are prioritizing captioned content in their algorithms, and users increasingly expect text alternatives for every piece of audio and video content they consume.
This guide covers why transcription matters for accessibility, what the law actually requires, and how to implement captions efficiently using modern AI tools.
The accessibility case for captions
Who benefits from captions
Captions are often framed as a feature for deaf and hard-of-hearing users, but the actual beneficiary list is much broader:
- Deaf and hard-of-hearing viewers (approximately 430 million people globally with disabling hearing loss, according to the WHO)
- Non-native speakers who understand written language better than spoken, especially at natural speaking speeds
- Viewers in sound-off environments such as offices, public transit, hospitals, and libraries
- People with cognitive or learning differences including ADHD, dyslexia, and auditory processing disorders, who often retain information better when they can read along
- Search engines and AI systems that cannot watch or listen to video but can index text transcripts
A study by Verizon Media and Publicis Media found that 80% of people who use captions are not deaf or hard of hearing. They use captions because it improves comprehension, allows watching in quiet environments, or helps with content in accented or fast speech.
Captions improve engagement metrics
Beyond accessibility, captions have measurable effects on content performance:
- View time: Facebook reported that captioned video ads increased view time by an average of 12%
- Comprehension: Multiple studies show 40-80% improvement in information retention when captions are present
- Reach: Captioned content is shareable to a wider audience, including the estimated 20% of global social media users who keep sound off by default
- SEO: Search engines index caption text, making captioned videos discoverable through text-based search queries. For more on how this works, see our article on the state of AI transcription in 2026
The engagement argument alone justifies captioning, even before considering legal requirements. For organizations building accessible content workflows, AI transcription has made compliance far more achievable.
Legal requirements in 2026
Accessibility legislation has expanded significantly in recent years. Here is where things stand.
United States
Americans with Disabilities Act (ADA)
Courts have consistently interpreted the ADA to cover digital content from businesses that serve the public. Multiple federal court decisions have ruled that websites and online video qualify as places of public accommodation. The practical effect: if your organization serves the public online, your video content should have captions.
Section 508
All federal agencies and organizations receiving federal funding must make electronic content accessible, including video. Section 508 references WCAG standards (see below) as the technical benchmark.
FCC regulations and the CVAA
The 21st Century Communications and Video Accessibility Act requires captions on internet video that was previously broadcast on television. The FCC enforces caption quality standards including accuracy, synchronicity, completeness, and placement.
European Union
European Accessibility Act (EAA)
The EAA took effect in 2025 and requires digital services, including video platforms and e-commerce sites, to meet accessibility standards. Member states are expected to enforce these requirements, and captioning is explicitly listed as a key component.
EN 301 549
This European standard for ICT accessibility references WCAG and includes specific requirements for captions and audio descriptions. It applies to public procurement and increasingly to private-sector digital services.
International standards
Web Content Accessibility Guidelines (WCAG) 2.1
WCAG is the de facto global standard for web accessibility, referenced by legislation in the US, EU, UK, Canada, Australia, and many other countries.
| WCAG level | Caption requirement |
|---|---|
| Level A | Captions for all prerecorded audio in synchronized media (SC 1.2.2) |
| Level AA | Captions for all live audio in synchronized media (SC 1.2.4) |
| Level AAA | Sign language interpretation for prerecorded content (SC 1.2.6) |
Most regulations require conformance to Level AA, which means captions for both prerecorded and live audio content.
The cost of non-compliance
ADA-related digital accessibility lawsuits in the United States have increased steadily, with thousands filed annually. Settlements and judgments often include requirements to remediate all existing content, implement ongoing accessibility programs, and pay damages. The legal costs of non-compliance frequently exceed what it would have cost to caption content proactively.
Beyond lawsuits, platforms like YouTube, Facebook, and LinkedIn are increasingly surfacing accessibility features in their algorithms. Uncaptioned content may receive less distribution than equivalent captioned content.
Captions vs transcripts: what you need
For full accessibility compliance, you typically need both captions and transcripts.
| Format | What it is | When to use |
|---|---|---|
| Closed captions | Time-synced text overlay on video, toggled by viewer | All video content |
| Open captions | Burned into the video frame, always visible | Social media, short-form content |
| Full transcript | Complete text document of audio content | Podcasts, audio-only content, supplementary resource |
| Audio description | Narrated description of visual elements for blind users | Video where visual information is essential to understanding |
WCAG Level A requires captions for prerecorded synchronized media. A transcript alone does not satisfy this requirement for video because it lacks time synchronization. However, for audio-only content like podcasts, a transcript is the standard accessible alternative.
The practical recommendation: provide closed captions for all video and a downloadable transcript as a supplementary resource. This covers the broadest range of accessibility needs and legal requirements.
For a deeper explanation of the differences between caption formats, see our guide on closed captions vs subtitles.
How to implement captions efficiently
Captioning used to be expensive and slow. Professional captioning services charge $1 to $3 per minute, and turnaround takes hours to days. AI transcription has changed the economics dramatically.
Step 1: Choose a transcription tool
Select a tool that supports your languages and export formats. For multilingual content or videos in languages other than English, language coverage is critical. Vocova supports over 100 transcription languages with automatic detection, which eliminates the need to manually specify the language for each video.
If you are evaluating tools, our best AI subtitle generators comparison covers the leading options.
Step 2: Transcribe your content
Upload your video or audio file, or paste a URL from platforms like YouTube, Vimeo, or Google Drive. AI transcription processes audio at many times real-time speed, meaning a one-hour video typically takes just a few minutes.
The output includes timestamped segments, automatic punctuation, and optionally speaker diarization to identify who said what. Speaker identification is particularly important for accessibility because it helps hearing-impaired viewers follow conversations.
Step 3: Review and edit
AI transcription is not perfect. Review the transcript for errors, especially:
- Proper nouns and brand names
- Technical terminology
- Acronyms and abbreviations
- Numbers, dates, and currency amounts
- Homophones (words that sound alike but have different meanings)
For accessibility captions, you should also add non-speech audio descriptions where relevant: [music playing], [applause], [phone ringing]. Current AI models focus on speech recognition and do not automatically annotate ambient sounds.
The word error rate of modern AI transcription on clean audio is typically below 5%, meaning most of your transcript will be correct. Focus your editing time on the error-prone categories listed above.
Step 4: Export in the right format
Export your transcript in the format your platform requires:
- SRT: The most widely supported subtitle format, works with YouTube, Vimeo, most video editors, and social platforms
- VTT: The HTML5 web standard, supports styling and positioning, required by some web players
- TXT: Plain text transcript for supplementary download or webpage embedding
- PDF/DOCX: Formatted documents for archival or distribution
For details on choosing between SRT and VTT, see our format comparison guide.
Step 5: Upload and verify
Upload your caption file to the video platform. Then verify:
- Captions are properly synchronized with the audio
- No segments are missing or out of order
- Speaker identification is correct
- Non-speech annotations appear at the right moments
- Caption display does not obstruct important visual elements
Step 6: Translate for multilingual accessibility
If your audience spans multiple languages, translate your captions to reach viewers who may need both hearing accessibility support and language support. Vocova supports translation into over 140 languages and can export bilingual captions with both the original and translated text.
Multilingual captions are not just a nice-to-have. For organizations operating internationally, they may be required under local accessibility laws that mandate content be accessible in the language of the jurisdiction.
Building an accessibility workflow
For organizations publishing video regularly, the key is making captioning part of the production process rather than an afterthought.
Integrate captioning into your publishing pipeline
Treat captions as a required deliverable, not an optional extra. Just as you would not publish a webpage without alt text on images, do not publish video without captions. Build captioning into your checklist:
- Record with good audio quality (see our guide on improving recording quality)
- Transcribe immediately after production
- Review and edit the transcript
- Export captions and transcript
- Upload captions with the video
- Verify synchronization and accuracy
Set quality standards
Define what "good enough" means for your captions:
- Accuracy target: Aim for at least 99% accuracy after editing. The FCC's caption quality standards require captions to be accurate, synchronous, complete, and properly placed.
- Turnaround time: AI transcription makes same-day captioning feasible for most content.
- Speaker identification: Required for multi-speaker content to maintain clarity.
- Non-speech annotations: Include for content where ambient sounds carry meaning.
Track compliance
Maintain an inventory of your video content and its captioning status. Identify gaps in your existing library and prioritize captioning based on traffic and audience reach. Most accessibility audits will check both new and existing content.
Frequently asked questions
Are captions legally required for all online video?
The legal requirements vary by jurisdiction and organization type. In the United States, the ADA has been broadly interpreted to cover online video from public-facing organizations. The EU's European Accessibility Act requires captions for digital services. WCAG Level AA, referenced by most regulations, requires captions for all prerecorded and live synchronized media. If you serve the public online, assume captions are required.
What is the difference between captions and a transcript?
Captions are time-synced text that appears on screen during video playback. A transcript is a standalone text document of the entire audio content. WCAG requires captions for video (time-synced) and transcripts for audio-only content. Providing both gives the most complete accessibility coverage. See our detailed captions vs subtitles guide for more.
How accurate do captions need to be?
The FCC requires captions to be "accurate," which courts and regulators have generally interpreted as 99% or higher accuracy. WCAG does not specify a percentage but requires captions to accurately represent the audio. AI-generated captions typically achieve 95-99% accuracy on clean audio, meaning light editing is usually needed to reach compliance standards.
Can AI-generated captions meet accessibility standards?
AI captions provide an excellent starting point and meet the core requirement of providing time-synced text for speech. However, for full compliance, you should review AI output for accuracy and add non-speech audio descriptions (sound effects, music cues, speaker identification) that current AI models do not generate automatically. The combination of AI transcription plus human review is the most cost-effective path to compliant captions.
How much does captioning cost with AI tools?
AI transcription tools range from free to approximately $0.05 to $0.10 per minute on paid plans. Vocova offers 30 free minutes to get started, Plus from $7.50 per month (billed annually), and unlimited Pro from $19 per month (billed annually). Compare this to professional human captioning services at $1 to $3 per minute. For a library of 100 hours of video, the difference is between roughly $50-100 with AI versus $6,000-18,000 with human services.
Do I need to caption old videos too?
If your organization is subject to accessibility requirements, existing content is typically included. Many settlement agreements require remediation of all published video content. Prioritize by traffic and visibility: caption your most-viewed and most-recent content first, then work backward through your library.
What about auto-generated captions on YouTube and Facebook?
Platform auto-captions are better than nothing but are not sufficient for compliance. They often contain errors, lack speaker identification, and do not include non-speech audio descriptions. The FCC and WCAG standards require accurate captions, and auto-generated captions frequently fall short. Use auto-captions as a starting point, but review and correct them before relying on them for accessibility.
