How to add subtitles to video lessons

Subtitles are one of those features that feel like extra work — until you look at the data. Studies from Netflix, YouTube, and Udemy all converge: videos with subtitles get 12–18% longer viewing time than videos without. In an online course, that translates directly into more completed lessons, happier buyers, fewer refunds.

On top of that: subtitles open your course to two audiences you'd otherwise lose — people who watch without sound (at work, on a commute, with a sleeping baby), and people with hearing impairments (~1.5% of any given European population).

In this guide I'll show you how to add subtitles to video lessons in audienced, which tools to use for automatic generation, and how to make sure they're genuinely good.

How subtitles work technically

audienced uses Bunny.net Stream as the video provider. The Bunny player supports:

  • SRT files (SubRip, the most widely used format).
  • VTT files (WebVTT, an advanced format with styling).
  • Multiple languages at once — the user picks in the player.
  • Auto-translate — via the Bunny Subtitle API (paid option).

Subtitles are a separate file from the video. You upload the MP4, then the SRT. The player syncs automatically.

Step 1: get a transcript (manual or automatic)

For subtitles you need a transcript with timestamps. Three paths:

A) Manual transcript

Most accurate, slowest. Open the video in Descript, Otter.ai or similar, write segment by segment, export as SRT.

Time: 3–4× video length (25-min video = 75–100 min of work).

B) AI transcription + manual cleanup

Recommended. Tools:

  • Whisper (OpenAI) — free, the best accuracy for most European languages. Run locally via CLI or online.
  • Descript — €15/month, great UX, timeline editing.
  • Rev.com — €0.25/min for an AI transcript.
  • Otter.ai — strong for English, mid-tier for smaller languages.

Time: 20 min of machine work + 30 min of your cleanup = about 50 min for a 25-min video.

C) Bunny.net auto-transcription

Right inside audienced — Bunny's API can generate subtitles automatically. In the video settings click Generate subtitles (auto) → pick the language.

Time: 5–10 min of machine work, no intervention from you.

Quality: excellent for English. For smaller European languages it's good but usually needs ~10–15% manual cleanup (names, technical terms).

Step 2: prepare the SRT file

An SRT file looks like this:

1
00:00:00,000 --> 00:00:03,500
Welcome to the first lesson.

2
00:00:03,600 --> 00:00:07,200
Today we'll look at what matters most
for success in the first week.

3
00:00:07,300 --> 00:00:10,000
Let's start with the question I get most often.

Rules for good subtitles

  • Max 2 lines on screen. More is hard to scan.
  • Max 42 characters per line. More gets cramped.
  • 1–7 seconds per segment. Under 1s is too fast, over 7s drags.
  • Break at natural pauses. Not mid-sentence if you can help it.
  • Don't split names or technical terms across two screens.

Step 3: upload subtitles in audienced

In the video lesson editor find Video → tab Subtitles.

Click Upload subtitles. For each file:

  • Pick the language (English, Slovenian, Croatian, German, Serbian, Czech).
  • Upload the SRT or VTT file.
  • Mark whether this language is default (turned on automatically for users with the matching profile language).

Click Save. Bunny processes the SRT, and it's available in the player within minutes.

Step 4: check and test

Open the video as a buyer would. In the player click the CC icon. Verify:

  • Subtitles appear at the right moment (sync).
  • Segments are split correctly.
  • Spelling is clean.
  • All languages you uploaded are visible.

Also check full-screen mode, mobile player, and different video speeds (0.75x, 1.25x). All must work.

Multiple languages — strategy

If your course audio is in your native language but you have buyers across Europe, I recommend:

  • Source language — main subtitles (matches audio).
  • English — translated subtitles for international reach.
  • One or two regional languages — for your closest neighbouring markets.

AI translation (ChatGPT, Claude, DeepL) will translate an SRT from language A to language B in 2 minutes. Quality is around 90% good; you need 20 minutes of native-speaker review.

That alone can bring in extra EU buyers without a new course — same MP4, extra SRTs, separate brand / landing page.

Subtitle styling (VTT)

VTT supports advanced styling:

  • Text colours.
  • Position on screen (top / bottom / middle).
  • Formatting (bold, italic).
  • Font size.

In audienced we currently use the default Bunny Stream player styling. Style customisation requires the enterprise plan.

Frequently asked questions

How long does it take to add subtitles to an entire course?

For a 10-lesson course with auto-transcription + manual cleanup: 3–5 hours total. The first lesson takes the longest (setting up the workflow), the rest are fast.

Do I need subtitles for every video?

Technically no, practically yes. Inconsistent subtitling (5 of 15 lessons) gives a bad user experience. Either all or none.

Do subtitles help SEO?

Inside the course, yes (users find specific lessons via Ctrl+F on the transcript). For Google SEO, video isn't indexed directly unless you publish the transcript on your blog as well.

What if I update the video — do subtitles adapt automatically?

No. If you swap the MP4, you have to regenerate the SRT (timestamps change).

Can I offer subtitles as a download to buyers?

Currently they're not available as a free download. You can include them as a PDF (transcript) inside module resources.

What does the Bunny auto-subtitle API cost?

~€0.01 per minute of video. A 25-min video costs ~€0.25. For 100 lessons, €25 one-off.

Does audienced support subtitles for live events (webinars)?

For recordings yes, for live streaming currently no (Bunny Live doesn't expose a real-time captions API).

Which language should the subtitles be in if I record in, say, Slovenian?

Always the source language first (match audio). Then add languages based on your target audience.

Closing thoughts

Subtitles are a small investment with a disproportionately large impact on user experience and retention. 3–5 hours of work for a full course pays off on the first 50 buyers; after that it's pure bonus.

If you're launching a new course, add subtitles from lesson one. If you have an existing course without subtitles, start with the 3 most important lessons (intro, core explanation, conclusion) and expand from there.

Try audienced 14 days free