WordPress Plugins for Video Captions, Transcripts and Audio Description (2026)

Media accessibility has an unusual property: the compliance requirement and the SEO benefit point in exactly the same direction. Captions and transcripts are required under WCAG for pre-recorded media, and a transcript is also the only version of your video that a search engine can read. Almost nothing else in accessibility pays for itself this directly.

The economics have also changed. AI transcription now runs roughly $0.05 to $0.25 a minute against $0.72 to $1.50 for human work, which means captioning an hour of video costs somewhere between $3 and $15 rather than $45 to $90. At those prices the argument for not captioning has essentially disappeared. This article covers what is actually required, prices the options, and is clear about where automatic output is not good enough.

Verified August 2026. Related: our voice and audio guide covers text-to-speech in the other direction.

Captions, subtitles, transcripts, audio description

Four things, routinely conflated, with different requirements attached.

Captions are synchronised text of the spoken audio plus relevant non-speech sound: a door slamming, laughter, music that carries meaning. They are for people who cannot hear the audio, and the non-speech part is what distinguishes them.
Subtitles assume you can hear but not understand, typically a translation. They omit sound effects. Subtitles do not satisfy a captions requirement.
Transcripts are the full text as a document, not synchronised. They serve people who prefer reading, people using braille displays, and search engines. For audio-only content a transcript is generally sufficient.
Audio description narrates important visual information the soundtrack does not convey. It is required for pre-recorded video at AA where visual content carries meaning the audio does not, and it is the expensive one.

The practical implication: auto-generated subtitles from a video platform are usually neither captions nor a transcript, because they omit non-speech sound and are not offered as a document. They are a starting point.

The 6 options compared

1. Rev

The option that spans both ends of the accuracy range, which makes it the easiest to reason about. Its AI transcription runs $0.25 a minute, around $15 an hour, while human transcription at roughly 99% accuracy is about $1.25 a minute, with professional captions around $1 a minute. AI subscription plans start near $29.99 per seat per month.

Having both under one account is genuinely useful, because the correct strategy is usually mixed: AI-transcribe everything, then send your highest-value or most technical content for human work. A product demo full of proper nouns, or a legal briefing where a misheard word changes the meaning, justifies $1.25 a minute. A weekly team update does not. Rev also delivers caption files in the formats video players expect, which removes a conversion step.

Pricing: AI $0.25 a minute; human transcription ~$1.25; captions ~$1
Strength: AI and human under one account
Best for: Mixing automatic and professional by content value
Watch out for: Seat-based subscription on top for AI plans

2. Happy Scribe

Subscription plans running $17 to $49 a month, dropping to around $8.50 a month on annual billing for the entry tier, with a human review service at $2 a minute for 99% accuracy. Strong multilingual coverage is its distinguishing feature, which matters if you publish in several languages and want captions in each.

Its human review is priced above Rev’s, so if you need substantial human work Rev is cheaper. Where Happy Scribe wins is the subscription model for steady volume: a site publishing a few hours of video a month gets predictable costs rather than a per-minute bill that varies. The annual discount is significant enough that monthly billing is worth avoiding if you know you will keep using it.

Pricing: $17 to $49 a month; ~$8.50 annually on entry; human review $2 a minute
Strength: Multilingual coverage and predictable subscription cost
Best for: Steady publishing volume across languages
Watch out for: Human review costs more than Rev’s

3. Otter.ai

Basic at $17 a month, Pro at $29, and Business at $89, with Pro dropping to around $8.49 a month billed annually. Built primarily around meetings rather than published media, which shapes both its strengths and its limits.

Speaker identification is its standout capability, and for anything with multiple voices, a panel discussion, an interview, a recorded webinar, that turns a wall of text into a usable transcript. If your published content is largely conversational, that is worth more than a small accuracy difference. Its caption file output is less central to the product than Rev’s or Happy Scribe’s, so check the export formats match what your player needs before committing to it as a captioning workflow.

Pricing: Basic $17, Pro $29, Business $89 a month; Pro ~$8.49 annually
Strength: Speaker identification on multi-voice recordings
Best for: Interviews, panels, and recorded webinars
Watch out for: Meeting-oriented; verify caption export formats

4. Descript

From $24 a month, or around $16 billed annually, and structurally different from everything else here: it treats the transcript as the editing interface, so deleting a sentence of text removes it from the video.

For accessibility that produces a useful side effect. Because you edit through the transcript, the transcript is necessarily accurate to the final cut, which removes the most common failure in captioning workflows: captions generated from an earlier version that no longer match what viewers hear. If you are both producing and captioning video, this collapses two jobs into one. If you only need captions for existing finished video, it is more tool than the task requires and the per-minute services will be cheaper.

Pricing: From $24 a month, around $16 billed annually
Strength: Transcript-based editing keeps captions matched to the cut
Best for: Teams producing and captioning their own video
Watch out for: Overkill for captioning existing finished footage

5. Platform auto-captions

Free, immediate, and the correct place to start rather than the place to stop. YouTube and Vimeo generate captions automatically on upload, and the quality is now good enough to be a usable first draft for clear speech in a common language.

What they will not do is meet the requirement unaided. Auto-captions omit non-speech sound, so they are subtitles rather than captions in the accessibility sense. They mangle proper nouns, technical terms, and product names reliably, which on a business site means they mangle exactly the words that matter. And accuracy collapses with accents, crosstalk, and background noise. The workflow that works is to let the platform generate, then spend ten minutes per video correcting names and adding sound cues. That is free and produces something genuinely compliant.

Pricing: Free
Strength: Immediate first draft at no cost
Best for: Starting point before human correction
Watch out for: Omits non-speech sound; mangles proper nouns

6. Presto Player and accessible players

The delivery layer, and a component people forget entirely. A perfectly captioned video in a player whose controls cannot be reached by keyboard is still inaccessible, and many custom-styled players fail exactly there.

What a player must do: expose play, pause, volume, and caption toggle to the keyboard; label those controls for screen readers; and let users change caption appearance. Presto Player is the common WordPress choice and integrates with the major hosts while keeping the video on a platform that handles delivery. The test is the same one that applies everywhere in accessibility: load a page with your video, unplug the mouse, and try to play it and turn captions on. If you cannot, the captions you paid for are unreachable.

Pricing: Free tier available; paid tiers for advanced features
Strength: Keyboard-reachable controls and caption toggling
Best for: Making the captions you produced actually usable
Watch out for: Custom-styled players frequently break keyboard access

Comparison table

Option	Cost	Accuracy tier	Caption files	Best for
Rev	$0.25/min AI; ~$1.25/min human	Both	Yes	Mixing AI and human by value
Happy Scribe	$17 to $49/mo; $2/min human	Both	Yes	Multilingual, steady volume
Otter.ai	$17 to $89/mo	AI	Check formats	Multi-speaker recordings
Descript	From $24/mo	AI	Yes	Producing and captioning together
Platform auto-captions	Free	AI, unreviewed	Yes	First draft only
Accessible player	Free tier available	Not applicable	Displays them	Making captions reachable

A workflow that is affordable and compliant

Auto-generate everything. At $0.05 to $0.25 a minute, or free on a platform, there is no content cheap enough to justify skipping this step.
Correct proper nouns first. Your product names, people’s names, and technical terms are what AI gets wrong and what your audience notices. Ten minutes per video.
Add non-speech sound cues. This is what turns subtitles into captions, and it is usually a handful of entries per video.
Publish the transcript on the page. Free, satisfies the transcript requirement, and gives search engines the only readable version of your video. This step pays for the whole exercise.
Send high-value content for human review. Anything legal, medical, technical, or heavily promoted justifies $1 to $2 a minute. Most content does not.
Test the player with a keyboard. Captions you cannot switch on are captions you did not deliver.

On audio description, be realistic. It is the most expensive requirement and it only applies where visuals carry meaning the audio does not. The cheapest fix is often production rather than post-production: narrate what you are showing while you record it, and the need for separate description largely disappears.

Related guides

Frequently asked questions

Are YouTube auto-captions enough?

No. They omit non-speech sound, which makes them subtitles rather than captions, and they misrender proper nouns and technical terms. As a first draft you then correct, they are excellent and free.

Do I need both captions and a transcript?

For video, captions are required and a transcript is strongly advisable. For audio-only content a transcript is generally sufficient. Publish the transcript regardless, because it is free once you have the captions and it is the only version search engines read.

What does audio description actually cost?

Considerably more than captions, because it requires scripting, voicing, and mixing. The cheaper route is designing it out: narrate the visuals as you record, and the separate description is often unnecessary.

Is AI accuracy good enough now?

For clear speech in a common language, close. It still fails on proper nouns, technical vocabulary, strong accents, and crosstalk. Budget correction time rather than assuming the output is publishable.

Do captions help SEO?

The transcript does, substantially, because it is the only text representation of your video that a crawler can index. This is the rare accessibility requirement with a direct commercial return.

What about live video?

Live captioning is a different and more demanding requirement, usually met with a real-time service or a live captioner. If you stream regularly, price it separately; the pre-recorded tools here do not cover it.

The verdict

For most sites: platform auto-captions to generate, ten minutes of correction per video for proper nouns and sound cues, and the transcript published on the page. That is free, compliant, and it improves your search visibility.

For volume: Rev at $0.25 a minute for AI, escalating to around $1 a minute for professional captions on your most important content. The mixed strategy is almost always the right one.

For multilingual publishing: Happy Scribe, from around $8.50 a month billed annually, with the strongest language coverage here.

And check your player with a keyboard. A captioned video in a player whose caption toggle cannot be reached is a requirement you paid for and did not meet.

ThemeAlley

Search

Recent Posts

Categories

Tag Cloud