Maybe it is a monthly podcast your firm records about tax changes. Maybe it is the pastor’s Sunday message posted as audio, a recorded board meeting, or a set of short audio tours for a museum. Someone who cannot hear it, or who cannot play sound at work, has no way in. The fix is old and simple: write it down.
Transcripts for accessibility are the text version of what was said and heard, placed where people can find it. This post explains when you need one, what a transcript should include, where it should live on your site, and what audio description is, in brief. Captions for video have their own post; this one is about the words that sit on the page.
When transcripts for accessibility are needed
WCAG treats media in three broad shapes, and the answer differs for each.
| Media | Examples | What people need |
|---|---|---|
| Audio only | Podcasts, recorded talks, sermons, voicemail-style updates, audio tours | A text transcript. At Level A, WCAG expects a text alternative for prerecorded audio-only content. |
| Video only, no sound | A silent screen recording, a product spin, a time-lapse | A text description of what happens, or an audio track that describes it. |
| Video with sound | Testimonials, webinars, explainers | Captions first. A transcript is a strong addition, and it is how many deafblind visitors, reading on braille displays, get the content. |
So for a podcast or any other audio file on your site, the transcript is not an extra. It is the accessible version. For video, captions do the main job, and the transcript serves people who want to read rather than watch, search for a quote, or skim a long recording. Our post on captioning website videos covers the caption side.
Transcripts help far more people than the ones WCAG was written for. People in noisy places, people with slow connections, people reading in a second language, and people who simply read faster than they listen all use them.
What a useful transcript includes
A transcript is more than a dump of speech-to-text output. A good one includes:
- Every spoken word that carries meaning, with names, numbers and terms spelled correctly.
- Speaker labels, so the reader knows who said what. Use names on first mention (“Dana Ruiz, host”) and short labels after that.
- Meaningful sounds in square brackets: [audience laughs], [phone rings], [music fades].
- Visual information, for video transcripts: what appears on screen that the speech does not say, like a slide’s headline or a demonstration.
- Links to anything mentioned, such as a form, a guide or a guest’s site, written as real links in the text.
- A title and date so the transcript matches the right episode or recording.
Timestamps are optional. They help on long recordings where people may want to jump to a section, but a timestamp on every line makes reading harder.
Verbatim or cleaned up?
Full verbatim keeps every “um”, false start and repeated word. It suits legal or research uses. For most business and nonprofit content, a clean verbatim style works better: remove filler and stumbles, keep the speaker’s words and meaning, do not rewrite their sentences. If you clean up, do not change facts, soften statements or add things that were not said.
Making podcast transcripts without a big budget of time
Podcast transcripts are where most small organizations fall behind, because episodes keep coming. A workflow that keeps up:
- Generate a draft with a speech recognition tool. Many recording and editing tools offer this, as do hosting services.
- Correct names, numbers, product names and anything that changes meaning. Listen at normal or slightly faster speed with the text in front of you.
- Add speaker labels and break the text into paragraphs when the speaker changes or the topic shifts.
- Add sound notes and link anything mentioned.
- Publish the transcript at the same time as the episode, not “later”. Later tends to mean never.
Keep a short glossary for your show: guest names, local place names, your products, recurring terms. Feeding it to the tool, or just checking against it, saves time on every episode.
If your back catalog is large, start with the episodes people still find through search or that you link from other pages, then work backward.
Where to put the transcript
A transcript only helps if people can find it without hunting. In order of preference:
- On the same page, right below the player, under a heading such as “Transcript”. Screen reader users can jump to it by heading, and everyone else can scroll.
- In an expandable section below the player, if the transcript is long. The control that opens it must be a real button that works with a keyboard and tells screen readers whether it is open or closed. Some accordion widgets get this wrong, so test it with Tab and Enter.
- On a separate page, linked right next to the player with clear text like “Read the transcript of episode 12”. This works, but it is one more step.
Things to avoid: a transcript only inside the podcast app, where visitors to your site never see it; a transcript only as a download; and a transcript as a scanned or untagged PDF. A transcript is text, and the web page is the natural home for it. If a document must be offered too, our guide to accessible PDFs for websites explains what makes one usable.
Formatting the transcript page
Treat a long transcript like any other long page. Use headings for segments or topics, a new paragraph for each speaker turn, and the speaker label in bold at the start. Avoid one huge block of text: it is hard to read for everyone and very hard for people with cognitive or reading disabilities. If there is a show summary, put it above the transcript so the reader knows what they are getting.
Audio description, in brief
Transcripts and captions turn sound into text. Audio description goes the other way: it turns important visuals into speech, for people who are blind or have low vision. Think of a video where someone says “as you can see here” while pointing at a chart, or a product demo where the steps are shown but never spoken.
WCAG has criteria at Level A and AA about making that visual information available in prerecorded video. For a typical business video, the simplest way to meet them is in the script, before filming: have the speaker say what is on screen. “This chart shows wait times dropping from March to June” needs no extra description track. Retrofitting description onto an old video means writing descriptions to fit the pauses, or producing an alternative version, which is more work.
A descriptive transcript, which combines the speech with written descriptions of the important visuals, is a practical companion for many videos. It is the same idea as a transcript with the visual notes filled in properly.
Checking your transcripts
A quick review before you publish:
- Read the transcript without listening. Does it make sense on its own?
- Check three or four random spots against the audio for accuracy.
- Confirm every speaker is labeled and every link works.
- Find the transcript from the player using only a keyboard.
- Zoom to 200 percent and check the transcript still reads in one column without sideways scrolling.
Writing transcripts for accessibility belongs with the other content habits in our guide to accessible web content for editors. If you have a long archive of audio and video that needs text, our captions, transcripts and media accessibility page explains how that cleanup can be organized. Our accessibility services for US businesses handle the rest of the site, from audits to monthly checks.
Frequently asked questions
Do podcasts legally need transcripts?
We cannot give legal advice, and the answer depends on who you are and which law applies. In accessibility terms, WCAG at Level A expects a text alternative for prerecorded audio, and WCAG is the benchmark most US audits and demand letters refer to. If you have a specific legal question, ask a lawyer; the technical answer is to publish a transcript.
Is an AI-generated transcript enough?
It is a good first draft. Speech recognition gets most ordinary speech right but still mishears names, numbers and specialist terms, and it often misses speaker changes and sounds. Someone needs to read it through and correct it against the audio before it is published. An uncorrected transcript with wrong facts can mislead the very people relying on it.
Should the transcript be hidden behind a toggle to keep the page tidy?
That is fine if the toggle is a real button that works with a keyboard and announces whether it is open or closed. Label it clearly, for example Show transcript, and place it right below the player. If the widget cannot be operated without a mouse, put the transcript in plain view instead.
Are show notes or an episode summary enough instead of a transcript?
No. Show notes tell people what an episode is about, but a transcript gives them what was actually said, which is what a listener gets. Keep the summary, since it helps everyone decide whether to read on, and place it above the full transcript rather than in place of it.
Transcripts are easy to add on a well-built site. If your theme’s audio player is hard to use with a keyboard, or your page templates make long text hard to read, it may be worth looking at accessible website redesign, so media pages are built for every visitor from the start.
Next in this series: Accessible PDF Files on Your Website: What to Fix
