Captions should fit into the edit you are already making. SmoothyApp uses local Whisper transcription to turn your Premiere timeline or a media file into editable subtitles. No SmoothyEdit account is required.
The caption workspace brings the source, model settings and editable cues together. You work with the words and timings, then review their appearance in your video editor.
Start with the source you have
Open Captions in SmoothyApp. Use a Premiere sequence with the bundled panel connected, or select a supported audio or video file from your computer.
You can also drop one media file into the Captions panel. Choosing or dropping a file prepares the input; transcription starts when you choose Generate. A canceled picker or invalid drop leaves your previous selection available.
You do not need Premiere for the local-file workflow. Generate captions from a recording and export an SRT for an editor that accepts subtitle files. Direct Premiere import needs the desktop app, the bundled panel and an open project.
Choose a language and model
Whisper models run on your computer after an initial download. Smaller models need fewer resources; larger models need more memory and processing time. The useful choice depends on your recording and hardware.
Choose the spoken language or use automatic detection. Listen to a representative part of the recording and review the text before processing becomes part of your normal workflow.
For a complete Arabic workflow, see Arabic captions for Premiere Pro. Arabic and other supported languages are outside Premiere’s native transcription list; the exported subtitles can still be imported into Premiere.
For Arabic and dialect speech, consider Arabic + Large v3 Turbo. Tiny and Base can miss words or repeat phrases, especially with dialects. A larger model can help, but names, music and overlapping voices still deserve a careful check.
Review the words and timing
A generated transcript is a draft. Read it while listening to the source, correct names and phrases, and adjust cue timings when the words arrive too early or linger too long.
Use transcript-wide find and replace for repeated corrections. Smoothy lets you undo a replacement in one step. Caption editors use automatic writing direction for the text they contain.
Review captions against speech rather than treating another automatic transcript as ground truth. Two systems agreeing does not establish that either one heard a word correctly.
Reformat without transcribing again
Change the line length, line count and duration settings to suit your output. Reformatting uses the transcript you already have, so you can try another layout without rerunning Whisper.
Read the result at the size it will appear in your video. Break long thoughts into comfortable cues, and check that captions remain readable through fast exchanges.
Export the subtitle file when it is ready, or import captions into Premiere. DaVinci Resolve Free accepts the exported SRT. Generate from a matching local audio/video file, save the subtitles and import them into a Resolve subtitle track. No Premiere or paid Resolve upgrade is needed for this file workflow.
Keep the caption workflow local
Download a model once, then transcribe and edit without sending your recording to a cloud provider. Caption generation, editing, reformatting and subtitle export are free local desktop tools.
Optional Studio cloud tools have a separate account and credit system. You do not need them to make captions. Explore the caption page or download the free app.