Learn about captions
Captions and subtitles are time-synchronized texts that help your viewers understand the audio components of your video content.
| Synchronized Text | Description |
|---|---|
| Captions | Display the text of the original audio and include non-speech elements like noises |
| Subtitles | Translate the text of the original audio into the preferred language of the viewer Non-speech elements are not captured. |
For simplicity, our documentation refers to both captions and subtitles as captions.

Screenshot of an English caption in a video
When you upload a video or perform a live-to-VOD conversion, the JW Platform automatically extracts any caption tracks present in the original media. For new hosted uploads, JWX auto-generates a caption track from the video’s audio. You can also manually upload additional captions after ingest. Up to 50 caption tracks can be associated with a media item.
You can manage and edit all caption tracks and their metadata through the dashboard or API.
Benefits
When added to your video content, captions benefit you and your viewers.
| Publishers / Broadcasters / Creators | Viewers |
|---|---|
|
|
Features
| Feature | Description |
|---|---|
| Track Extraction | Captions are automatically extracted from the media file during upload or live-to-VOD conversion. |
| Metadata Retrieval & Editing | Captions include label and language metadata (if available). Defaults to en if not present. NOTE: Caption tracks follow the order of the uploaded file. Manually added tracks stay where they are placed. |
| Additional Track Management | Add new caption files after upload to supplement or replace auto-extracted captions. |
| Auto-generation | Captions are automatically generated from the audio of new hosted uploads at no additional cost. |
Format and delivery
Captions can be embedded in the video or side-loaded from a separate file.
| Format | Type | Delivered As | Description |
|---|---|---|---|
| CEA-608 CEA-708 |
Embedded | WebVTT | Standard formats originally designed for analog and digital television broadcasts Style is based on the viewer’s device settings when delivered in live streams. |
| SubRip Text (.srt) | Side-loaded | SRT | Simple caption file without styling or metadata |
| TTML (.ttml) | Embedded | WebVTT | Industry-standard, XML-based caption stream. JWX **only** supports TTML files that are embedded in the original media upload. TTML caption files **cannot** be created separately or uploaded to existing media. |
| WebVTT (.vtt) | Side-loaded | WebVTT | Common online standard format. Supports descriptions and metadata. Header must contain a WEBVTT string, and the file must be UTF-8 encoded. formatting a .vtt file |
Caption types
The JW Platform manages several types of captions associated with your media.
By default, all caption tracks in the original file appear in the player's caption options. To hide a track, you must delete it.
| Feature | Description |
|---|---|
| Auto-extracted text tracks | Text tracks extracted from the original VOD media or live-to-VOD conversion Track labels are prefixed with “[auto-extracted]” and saved in draft status ( auto_publish=false). Supported formats:
A `track_kind` of captions is assigned by default. |
| Auto-generated text tracks | Text tracks generated from the primary audio track of new hosted uploads Auto-generated tracks are saved in draft status unless the property automatically publishes auto-generated captions. Learn more. |
| Manually uploaded or created text tracks | Caption files uploaded via the dashboard or API, or captions created in the dashboard after ingest Supported formats:
|
Auto-generated captions
JWX automatically generates a caption track for new hosted uploads. Auto-generated captions are available to all customers at no additional cost.
Auto-generated caption tracks are saved in draft status (with the prefix [auto-generated]) and do not appear in the player until you publish them. You can choose to publish auto-generated captions automatically.
Supported media
| Media type | Auto-generated captions |
|---|---|
| Hosted uploads | Supported |
| URL-registered media | Not supported |
| CMAF VOD media | Not supported |
How auto-generated captions work
| Item | Description |
|---|---|
| Audio source | Only the primary (default) audio track is captioned. |
| Number of tracks | One auto-generated caption track is created per upload. |
| Language | The caption language is determined by the language detected in the audio. |
| Existing caption tracks | Auto-generated captions are added alongside existing caption tracks, including published tracks. |
| Reuploads | Reuploading a video replaces any earlier auto-generated caption tracks. Caption tracks that you added manually are kept. |
Publish auto-generated captions automatically
Follow these steps to publish auto-generated captions automatically for all new uploads to a property:
- From the Properties tab, click the name of the property. The property settings appear.
- Under Content & metadata, click Media enhancements. The Media enhancements tab appears.
- Under Auto-generated captions, toggle Publish automatically to ON.
When toggled OFF, auto-generated caption tracks are saved as unpublished drafts. This is the property’s default setting.
FAQ
What is the targeted accuracy level of captions?
JWX uses the latest technology to generate captions, but accuracy may vary. Review captions before publishing.
If you require a guaranteed level of accuracy, contact your JWX account team to discuss available options.
Updated 7 days ago
