How to add subtitles to Bunny Stream videos

Bunny Stream takes caption files in two formats: .vtt and .srt. A video can have one caption track for each language. There are three ways to get a track onto a video.

Where captions are in the dashboard

On 2026.10.13, the dashboard path to your library was Delivery › Stream, then your library. The dashboard page of the Bunny docs lists a Transcribing and accessibility area in the library. It covers automatic captions and translation.

The docs that we checked do not list a dashboard step to upload a caption file by hand. They list the API for that. Open a video in your library and look for a captions option. If you do not see one, use the API steps below.

The Add Caption endpoint sends one caption file for one language to one video. The file goes in the JSON body as Base64 text.

1. Open your library in Stream and click API. Copy the library ID and the Stream API key. Every library has its own key.
2. Find the video's GUID. The Stream API uses it as the video ID.
3. Save the captions as UTF-8 text in a .vtt or .srt file. Check the file with the SRT validator first.
4. Convert the file to Base64 text with no line breaks.
5. Send a POST request to https://video.bunnycdn.com/library/{libraryId}/videos/{videoId}/captions/{srclang}. Put the language code in the path, for example es. Send your API key in the AccessKey header.
6. Put label and captionsFile in the JSON body. The label is the name for the track. The captionsFile field holds the Base64 text.
7. Read the answer. Status 200 means Bunny stored the track. Status 400 means the upload failed. The answer then has an errorList and a warningList that tell you what is wrong with the file.

1LIBRARY_ID=1234562VIDEO_ID=00000000-0000-0000-0000-0000000000003CAPTIONS=$(base64 < captions-es.vtt | tr -d '\n')45curl --request POST \6  --url "https://video.bunnycdn.com/library/$LIBRARY_ID/videos/$VIDEO_ID/captions/es" \7  --header "AccessKey: $BUNNY_STREAM_API_KEY" \8  --header "Content-Type: application/json" \9  --data "{\"label\":\"Español\",\"captionsFile\":\"$CAPTIONS\"}"

• The language comes from the srclang part of the path. The srclang field in the body is deprecated, and Bunny ignores it.
• Send the call again for each language. Each language has its own path.
• To remove a track, send a DELETE request to the same path.

Not sure which file to send? Read Subtitle formats explained. To change an SRT file to VTT, use the free tool below.

Bunny calls this feature Transcribing. It is built on Whisper, a speech recognition model. It makes a caption track from the speech in the video. It can also translate that track. The docs list 56 languages for both, with two-letter codes such as en, es, and ja.

Turn it on for a library

1. Go to Delivery › Stream and open your library.
2. Open the transcribing settings in the Transcribing and accessibility area.
3. Turn on transcribing and choose the languages.
4. Upload videos. Bunny makes captions for each new video. Videos that you uploaded before you turned it on get nothing.

Run it for one video

Send a POST request to the video's /transcribe path in the Stream API. The body can set the source language and a list of target languages. This request overrides the library setting, and you pay for it even when the library setting is off. Add force=true to run it again on a video that Bunny already transcribed. Bunny bills the run again.

The Bunny docs also describe a dashboard route through the Smart Chapters tool: open the video, open the Chapters tab, and choose Generate Smart Chapters. If the video has no caption track, Bunny runs Transcribe to make one.

Price

Bunny charges $0.10 for each language and each minute of video. A 60-minute video in English and Spanish costs 60 × 2 × $0.10, which is $12. Bunny's pricing docs list no free allowance for transcribing. Bunny bills by use, with a $1 monthly minimum.

Check the language first

Bunny finds the spoken language from the first 30 seconds of audio. If that part has music or silence, the guess can be wrong. In that case, set the language yourself when you start the run.

When Bunny's captions are enough

For one language, with clear speech and a video that is not for a client, the automatic track is a good start and costs little. Watch the video once with captions on. Fix what is wrong before you publish. Bunny's docs describe no editor for the text, so a fix means a new file (see the limits below).

One video holds one track for each language code. Add a track for each language that you need. The Bunny player has a captions control for the tracks, and a separate menu for audio tracks.

• A viewer can pick a track in the player. The C key turns captions on and off.
• The player saves the viewer's choice in the browser. A viewer who picks Spanish once gets Spanish on the next visit.
• The docs for the player do not describe the layout of the captions menu. Test it in your own embed.
• Automatic tracks use the language code with -auto added, for example en-auto. The docs do not say how Bunny names translated tracks.

Bunny Stream also supports several audio tracks in one video. That is a different feature. It is for dubbing, not for subtitles.

The Add Caption answer has a valid flag and lists errors and warnings. It is not a quality check. For line length and reading speed, use the subtitle QC checklist.

The built-in ways work well for one language, or for a quick draft. They stop short in these cases:

• Review. The docs describe no way to assign a translator, comment on a line, or approve a track before it goes live.
• Fixes. The docs describe no text or timing editor. You fix the file elsewhere and send it again.
• Quality in other languages. The docs give no accuracy figure for the captions or the translation. A client who pays for subtitles may want a human to check each language.
• Cost across languages. Each extra language adds $0.10 per minute, again for every video.
• Many videos. Each upload is one API call for each language and video. You need a script to do it at scale.

CaptionsHero is a subtitle workflow tool for teams. It has no Bunny Stream integration. You make and approve the subtitles in CaptionsHero, download a file, and upload it to Bunny Stream with the API steps above.

1. Upload the video to CaptionsHero.
2. Make the subtitles. Use the AI draft, import an SRT or VTT file, or write them in the editor. The editor has a waveform timeline and Netflix-style QC.
3. Add the other languages. Use AI translation or a human translator. Each language is a separate track.
4. On Team and Studio, assign an editor and a reviewer to each language. The reviewer comments on exact lines and asks for changes.
5. Approve each track when the review is done.
6. Download the approved version as VTT or SRT. Do this for each language.
7. Upload each file to Bunny Stream with the Add Caption endpoint. Use the right language code in the path.

Later edits in CaptionsHero do not reach Bunny Stream. After a change, download the file again and send it with the same language code. Bunny replaces the old track.

Do you host video in other places too? See the guides for Vimeo, Wistia, and Mux. For Cloudflare Stream, which the Studio plan can publish to directly, see the Cloudflare Stream page.

The price is flat for each organization. There is no sales call. CaptionsHero imports and exports SRT and VTT only. Bunny Stream takes both, so no conversion is needed.