Captions and subtitles

Captions help viewers watch without sound, improve accessibility, and make short-form videos easier to follow.

Open the Captions tab

1. Open a project with speech or audio.
2. Click Captions in the left asset panel.
3. Choose whether to generate captions or import subtitle files.

Generate captions from timeline audio

1. Put your video or audio on the timeline.
2. Make sure the spoken section is audible.
3. Open Captions.
4. Choose a language or use Auto.
5. Click Generate captions.
6. Wait for progress states to complete.
7. Review generated caption clips on the timeline.

Caption generation can include these progress stages:

• Audio extraction.
• Model loading.
• Transcription.
• Caption generation.
• Text track insertion.

On-device transcription

The on-device path runs locally in the browser where configured.

Use it when:

• You want a more private workflow.
• You can wait longer for processing.
• You want to avoid sending audio to a hosted transcription provider.

The first run may take longer because the model needs to load and cache.

Hosted transcription

Some deployments can use hosted transcription.

Use it when:

• You need faster transcription.
• Accuracy matters more than local-only processing.
• Your workspace is configured for hosted transcription.

If hosted transcription is unavailable, use the local path or ask your workspace admin to configure it.

Language selection

Options can include:

• Auto — detect language automatically.
• A specific language — best when detection is unreliable.

Choose a specific language when the audio has accents, music, or mixed speech that confuses detection.

Insert generated captions

Generated captions are inserted as text elements on a text track.

After generation:

1. Move the playhead to the first caption.
2. Press Space to preview.
3. Select a caption clip.
4. Open Text properties to edit style.
5. Trim or move caption clips if timing needs work.

Import SRT files

1. Open Captions.
2. Click Import.
3. Choose an .srt file.
4. Confirm import settings.
5. Review the inserted caption clips.

SRT is a simple subtitle format with numbered cues and timestamps.

Import ASS files

1. Open Captions.
2. Click Import.
3. Choose an .ass file.
4. Confirm import.
5. Review styling and timing.

ASS supports advanced subtitle styling. Some styling may be simplified in the editor, so always preview after import.

Handle malformed subtitle cues

If the import finds malformed subtitle cues, the editor may show warnings.

When that happens:

1. Read which cues failed.
2. Import the valid cues if available.
3. Fix the subtitle file in a text editor if needed.
4. Re-import.

Common subtitle issues include overlapping timestamps, missing end times, invalid numbering, and unsupported style syntax.

Edit caption text

1. Select a caption clip on the timeline.
2. Open the right properties panel.
3. Go to Text.
4. Edit the words.
5. Preview the sentence in context.

Correct names, product terms, and technical words manually.

Adjust caption timing

Use the timeline:

• Drag the caption clip to move it earlier or later.
• Trim the left edge to change when it appears.
• Trim the right edge to change when it disappears.
• Split long captions with S.
• Delete duplicate or incorrect captions.

Preview with audio after every timing change.

Style captions manually

Caption elements use the same text controls as normal text overlays.

You can adjust:

• Font family.
• Font size.
• Color or gradient fill.
• Alignment.
• Background box.
• Stroke.
• Shadow.
• Letter spacing.
• Line height.
• Text animation.

For readability, use strong contrast and keep captions inside safe areas.

Use caption skins

Caption skins are prebuilt styles often applied by the agent.

Examples:

• TikTok bold.
• Minimal clean.
• Boxed contrast.
• Editorial highlight.
• Neon pop.
• Documentary lower-third.

See Caption skins for style details.

Caption diagnostics

Caption diagnostics help explain transcription problems.

They can point to:

• Missing audio.
• Very quiet audio.
• Music overpowering speech.
• Unsupported language choice.
• Malformed subtitle files.
• Empty transcript sections.

Use diagnostics before regenerating captions.

Troubleshooting

See also

• Text overlays and titles
• Auto-caption tutorial
• Caption skins
• Audio editing