Audio-only and Video-only (Prerecorded)

What it asks

Two single-channel media types are covered. Audio-only files (podcasts, recorded interviews, MP3 clips) must have a text transcript. Video-only files (silent product demos, animated GIFs that convey information, sign-only clips, b-roll explainers) must have either a text alternative or a synchronized audio track that conveys the same information. The point: a deaf user gets the podcast, a blind user gets the silent demo.

How to meet it

• Publish a full text transcript directly below the audio player, not behind a download link only.
• For silent video, write a text description that captures every action and on-screen text, placed on the same page.
• Alternatively, add an audio narration track to silent video that describes what is happening as it happens.
• Use machine transcription as a starting point only — Whisper / Otter output must be edited for accuracy and speaker labels.
• Mark transcript headings, speaker names, and timestamps with real HTML structure (<h2>, <ol>, <time>) — not a wall of <br> tags.
• For embedded social-media audio (Twitter Spaces, voice notes), provide the transcript in the surrounding page text.

Common failures

• Podcast episode pages with only show notes — the conversation itself is never transcribed.
• Auto-generated YouTube transcript that is 80% accurate, with proper nouns mangled and speaker boundaries missing.
• Silent product GIFs with no description anywhere on the page, common on marketing pages.
• Transcripts hidden behind “Click to expand” with no programmatic association to the player.
• Audio interviews where one speaker is identified but the other is “Speaker 2” throughout.

Why it matters

Transcripts are also the cheapest SEO win on any media-heavy site — search engines index them, sighted users skim them, and they unlock the content for deaf users in one move. Skipping transcripts is almost always a workflow problem, not a technical one.