WCAG Compliance With Transcripts: SC 1.2 Mapped (2026)

This is not legal advice. WCAG criteria are technical standards, not legal instruments. Whether a specific regulation requires conformance, and at what level, depends on your jurisdiction, your organization type, and the law in question. Consult legal counsel for compliance determinations.

The Criteria That Involve Transcripts

Transcripts are the most versatile accessibility artifact for audio content, but they cover only part of the WCAG 1.2 landscape. SC 1.2.1 requires a transcript for audio-only content. SC 1.2.3 lets you use a transcript as the media alternative for video at Level A. Beyond those two, transcripts support the work but do not substitute for captions or audio descriptions.

WCAG 2.2 left Guideline 1.2 entirely unchanged from 2.1. The criteria below apply equally under both versions. The new criteria in WCAG 2.2 address keyboard navigation, touch targets, and cognitive accessibility, not time-based media.

How SC 1.2 Is Organized

The nine success criteria in Guideline 1.2 map to three conformance levels:

Most regulations target Level AA. ADA Title II in the US requires WCAG 2.1 AA for state and local governments (compliance dates extended to April 2027 and April 2028 depending on population size). Section 508 for US federal agencies requires WCAG 2.0 AA. The EU European Accessibility Act, enforced from June 2025 under EN 301 549, currently references WCAG 2.1 AA. For a deeper comparison of these frameworks by country, see accessibility laws by country.

SC 1.2.1: Audio-only and Video-only (Pre-recorded), Level A

For audio-only content, provide a text transcript. For video-only content (no sound), provide either a transcript or an audio track describing what is shown.

A podcast episode, recorded lecture without video, or any standalone audio file falls under this criterion. The transcript must convey the same information as the audio: speech, relevant non-speech sounds, and speaker identification when multiple speakers are present.

What counts as sufficient: the W3C specifies "correctly sequenced text descriptions of the time-based visual and auditory information." A summary does not qualify. A transcript that skips sections or omits key content does not qualify. Slight word variations and misspelled proper nouns are generally acceptable when the meaning is preserved.

The transcript does not have to be embedded on the same page as the audio, but it should be easily reachable from it. HTML on-page or linked HTML both satisfy the criterion. A PDF transcript can satisfy the requirement but creates additional accessibility burdens; HTML is the cleaner path.

One exception: if the audio-only content is itself a media alternative for text content already on the page, and is clearly labeled as such, SC 1.2.1 does not apply.

If you just need a clean transcript for a podcast or recorded audio, ConvertAudioToText's audio-to-text tool produces a text transcript you can download, review, and publish next to your audio player.

SC 1.2.2: Captions (Pre-recorded), Level A

Every pre-recorded video with audio on a publicly accessible website needs synchronized captions.

"Synchronized" is the operative word. Captions must appear during the speech they represent, not before or after in bulk. They must also include relevant non-speech audio: "[door closes]", "[upbeat music]", speaker labels for multi-person content.

WCAG does not set a specific accuracy percentage. FCC and DOJ guidance points to 99% as the target for broadcast and public content. Accessibility professionals generally treat 95% as the practical floor. Auto-generated captions from video platforms typically fall short on proper names, technical terms, overlapping speakers, and non-speech audio.

The workflow for compliant captions:

1. Upload video directly to a transcription tool or extract the audio for a video-to-text conversion.
2. Generate a timed SRT or VTT caption file.
3. Run human review: fix names, technical terms, add non-speech audio cues, verify timing.
4. Upload the reviewed file to your video host as a manual caption track.

Auto-captions from your video host are a starting point, not an endpoint.

SC 1.2.3: Audio Description or Media Alternative (Pre-recorded), Level A

For pre-recorded synchronized video (video with audio), provide either audio description or a full text media alternative. You get to choose which.

This is the criterion most often misread. The "or" is real and load-bearing at Level A.

Option A, audio description: A separate narrator describes visual content during natural pauses in the dialogue. "She walks to the whiteboard and circles the third item on the list." This requires recording an additional audio track.

Option B, media alternative: A text document covering everything in the video, dialogue and visual content integrated. The person who cannot see or hear the video should be able to read this document and get the same information.

For talking-head videos, lectures, and presentations where visual content is relatively sparse, Option B is usually easier to produce. Start with a complete transcript from the audio, then add brief visual annotations during review.

For demonstration videos, product walkthroughs, or anything where the screen or action carries significant content not described in the dialogue, audio description tends to serve users better.

Note: if you choose Option B (the media alternative) for 1.2.3, you still need to satisfy 1.2.5 at Level AA, which requires audio description specifically.

SC 1.2.4: Captions (Live), Level AA

Live streams, webinars, and virtual events with audio need real-time captions.

Pre-recorded captions give you time for human review. Live captions do not. The main options:

CART (Communication Access Realtime Translation): A professional stenographer captions in real time. Accuracy is high. CART services typically run from around $60 per hour for English to $120-200 per hour depending on subject matter, provider, and duration. The gold standard for high-stakes live content.

Automatic Speech Recognition (ASR): Live ASR services from platforms like Zoom, Google Meet, Microsoft Teams, and others. Fast and low-cost, but accuracy drops with multiple speakers, accents, technical vocabulary, or audio quality issues.

Hybrid: ASR with a human reviewer monitoring and making corrections in real time. Balances cost and accuracy for most use cases.

For Level AA compliance, WCAG requires that captions be accurate and timely. CART reliably meets this bar. ASR alone typically meets it for controlled, single-speaker content with clear audio. Complex or high-stakes events warrant CART or hybrid coverage.

For live meeting transcription purposes, most teams use the built-in ASR of their conferencing platform during the event, then run the post-event recording through a more accurate transcription tool to produce the corrected, archived version.

SC 1.2.5: Audio Description (Pre-recorded), Level AA

At Level AA, audio description for pre-recorded video is required, not optional.

This closes the flexibility that 1.2.3 gives at Level A. You can no longer substitute a media alternative. If you provided a text alternative for 1.2.3, 1.2.5 adds a new requirement on top.

The practical path that satisfies both: provide audio description to meet 1.2.5. Doing so automatically satisfies 1.2.3 as well, since audio description is one of the two acceptable options there.

The practical AA combination for video content: captions (1.2.2) plus audio description (1.2.5). Many compliance workflows treat these as the Level AA pair.

A useful shortcut for content where pauses in dialogue are sufficient: during production, leave deliberate gaps in narration for description audio to be added. Retrofitting descriptions into dense, wall-to-wall dialogue is significantly harder than designing for it.

My take: for most educational and corporate video producers, the bigger challenge is not understanding the criteria but building the workflow. Treating captions and transcripts as part of the production process rather than post-production fixes cuts the cost in half.

SC 1.2.6: Sign Language (Pre-recorded), Level AAA

For pre-recorded video with audio, provide sign language interpretation.

Level AAA, rarely mandated by regulation. Most public content does not meet this, and conformance targets of AA do not require it. It is most relevant for content specifically targeting Deaf audiences or high-stakes communications in public services and healthcare.

SC 1.2.7: Extended Audio Description (Pre-recorded), Level AAA

When standard audio description is insufficient because there are not enough natural pauses in the dialogue, provide extended audio description: the video pauses to allow a longer description to be inserted, then resumes.

Level AAA. Rare in practice because most content has sufficient pauses. Where this applies: instructional content with continuous dense narration and critical visual information.

SC 1.2.8: Media Alternative (Pre-recorded), Level AAA

A complete text document covering all audio and visual content in a synchronized video, for any user who cannot access either the audio or the visual track.

This goes beyond captions or audio description alone. The document needs:

• All dialogue with speaker identification
• All relevant non-speech audio
• All significant visual content, including on-screen text, setting descriptions, and character actions
• Structural navigation (chapters, scene changes)

A well-prepared transcript provides the dialogue layer. The visual annotation work is human, but starting from a complete, speaker-labeled transcript reduces the time substantially.

SC 1.2.9: Audio-only (Live), Level AAA

Real-time text alternative for live audio-only content: live radio, audio conferences, audio-only broadcasts.

Level AAA, rarely achieved in practice. Real-time text alternatives for audio-only streams typically require live ASR or a live captioner. Very few live audio-only broadcasts meet this criterion.

A Practical AA Compliance Workflow

For a website with podcasts and recorded video targeting WCAG AA:

For podcast episodes (SC 1.2.1)

1. Transcribe the audio using an audio-to-text tool.
2. Human review: fix names, technical terms, verify speaker labels.
3. Publish the reviewed transcript as HTML on or near the episode page.

For recorded videos (SC 1.2.2 and 1.2.5)

1. Transcribe to get a timed caption file (SRT or VTT).
2. Human review: accuracy, speaker labels, non-speech audio cues.
3. Upload captions to your video host.
4. Record or commission audio description covering visual content during dialogue pauses.
5. Publish the audio description as an alternate audio track.

For live webinars (SC 1.2.4)

1. During the live event: use CART or ASR live captions.
2. After the event: run the recording through a transcription tool for the archived version.
3. Review and correct the archived captions before publishing with the recording.

Common Compliance Failures

Auto-captions published without review. Auto-generated captions miss proper names, technical vocabulary, speaker labels, and non-speech sounds. Review is not optional.

A summary instead of a transcript. SC 1.2.1 requires the same information as the audio, not an edited summary. Publishing "show notes" in place of a transcript does not satisfy the criterion.

No visual description for visual content. A product demo that shows the interface without narrating what is shown needs audio description or a media alternative. Captions alone do not cover visual-only information.

Transcript behind a download gate. A PDF download satisfies the letter of the criterion but creates access friction for screen reader users. HTML on the page or linked HTML is the cleaner path.

No speaker identification in multi-speaker content. When two or more people are speaking, both captions and transcripts need to identify who is saying what.

Documentation for Compliance Audits

When documenting your accessibility process:

• Name the target standard (WCAG 2.2 Level AA, or WCAG 2.1 AA if that is what your jurisdiction requires).
• Describe your generation process (AI transcription followed by human review).
• State which SC each artifact satisfies and for which content.
• Note known exceptions and remediation timelines.
• Reference the tools used for generation. The compliance belongs to your process, not to any single tool.

For more on the legal frameworks that reference these criteria, see accessibility laws by country and ADA compliance for audio content. For the user-experience dimension, transcripts for screen reader users covers how the artifacts actually get used.

Common Questions

Does a transcript satisfy all WCAG 1.2 requirements?

No. A transcript satisfies SC 1.2.1 for audio-only content and can satisfy the media-alternative option in SC 1.2.3, but it does not satisfy SC 1.2.2 (captions must be synchronized), SC 1.2.5 (requires actual audio description for video), or the live-content criteria.

What is the difference between SC 1.2.3 and SC 1.2.5?

SC 1.2.3 is Level A and gives you a choice: provide either audio description or a full text alternative for prerecorded video. SC 1.2.5 is Level AA and removes that choice, requiring audio description specifically. If you meet 1.2.5 with audio description, you automatically satisfy 1.2.3.

Do auto-generated captions from YouTube satisfy SC 1.2.2?

Not on their own. WCAG sets no specific accuracy percentage, but FCC and DOJ guidance points to 99% as the target, and industry practice treats 95% as a minimum floor. Auto-generated captions miss names, technical terms, speaker identification, and non-speech sounds. Human review is required before treating them as compliant.

Which regulations actually require WCAG 2.2?

As of mid-2026, most regulations still reference WCAG 2.1 AA or older. ADA Title II (US state and local governments) requires WCAG 2.1 AA. Section 508 (US federal) requires WCAG 2.0 AA. The EU EAA and current EN 301 549 reference WCAG 2.1 AA. WCAG 2.2 adoption into binding regulations is in progress.

Sources

• Web Content Accessibility Guidelines (WCAG) 2.2, W3C
• Understanding WCAG 2.2, W3C WAI
• Understanding SC 1.2.1: Audio-only and Video-only (Prerecorded), W3C WAI
• Understanding SC 1.2.5: Audio Description (Prerecorded), W3C WAI
• Making Audio and Video Media Accessible, W3C WAI
• ADA Title II Web Accessibility Compliance Deadline Extension, Federal Register April 2026
• Fact Sheet: ADA Title II Web Rule, ADA.gov