Create subtitle files from audio or video — local Whisper, no upload
Create Subtitle File is a local-first workflow for both audio and video. Right-click a source file in Explorer, choose your Whisper model and export format, and get an adjacent unique subtitle file without sending anything to the cloud. Video files are handled through automatic audio extraction before transcription, while audio files go straight into the same subtitle pipeline.
Why create subtitles locally?
No upload, no remote service
Speech is transcribed on your Windows machine. Your recordings, interviews, tutorials and clips never leave your disk.
One workflow for audio and video
The Explorer actions stay separate for convenience, but both routes converge to the same subtitle engine and output format.
Flexible subtitle output
Standard SRT is the default, with Advanced ASS and FrameShift subtitle-project output also available. Every export uses safe unique naming.
GPU when available, CPU when needed
DirectML accelerates Whisper on compatible Windows GPUs, with automatic CPU fallback if the GPU is unavailable or runs out of room.
Right-click simplicity
No editor to open, no project to create. The same Windows-first flow as the rest of FrameShift stays intact.
Built for long media too
The subtitle worker processes media in overlapping windows, merges boundary words cleanly, and rebuilds readable SRT cues.
Choose the Whisper model that fits
Whisper Base — Fast
A lighter local model for quicker downloads and faster runs when speed matters most.
Whisper Small — Recommended
The default choice in FrameShift. A balanced option for quality, speed and download size in everyday subtitle work.
Whisper Large-v3 Turbo — Quality
The largest option for users who want the highest accuracy and do not mind a heavier model download.
How it works
Right-click an audio or video file in Windows Explorer and choose FrameShift → Create Subtitle File.
Select your Whisper model — Base for speed, Small for the default balance, or Turbo for the highest accuracy.
Model preflight runs automatically — FrameShift downloads the required files on first use and verifies them before transcription starts.
Video adds one prep step — the first audio track is extracted automatically, then normalized into the same transcription input as audio files.
FrameShift writes the subtitle file locally as a clean adjacent
.srtoutput with unique naming and no source overwrite.
Add subtitles to video
Create Subtitle File stays focused on local transcription and export. Use the separate Add Subtitles to Video action when you want a selectable subtitle track or subtitles burned into the video.
Frequently asked questions
Can FrameShift create subtitle files from audio or video?
Yes. FrameShift transcribes both audio and video locally with Whisper AI and writes a uniquely named subtitle output next to the source.
What subtitle format does FrameShift output?
Standard SRT is the default. Advanced ASS and a FrameShift subtitle project are also available, and every output uses safe unique naming.
Does Create Subtitle File work for both audio and video files?
Yes. Audio and video appear as separate right-click actions in Explorer, but they share the same subtitle pipeline. For video files, FrameShift first extracts the audio track, then transcribes it into the same local SRT workflow.
Which Whisper models are available?
Three local Whisper models are available: Whisper Base for speed, Whisper Small as the recommended default, and Whisper Large-v3 Turbo for the highest accuracy. Models are downloaded on demand and verified before use.
Does subtitle creation require a GPU?
No. FrameShift uses DirectML GPU acceleration when available and automatically falls back to CPU if needed. The subtitle workflow stays fully local either way.
Can FrameShift burn subtitles into a video?
Yes. The separate Add Subtitles to Video action can add a selectable subtitle track or burn subtitles into a video. Create Subtitle File remains focused on local transcription and subtitle export.
Try Create Subtitle File in FrameShift 1.18.1
Keep subtitle creation local, simple and close to the source media.