Create useful text beside local media
A transcript can support review, captions, search, and notes. Local processing keeps the selected media on the device instead of sending it to a transcription provider.
Local CPU transcription
Turn a trusted local video into TXT, SRT, or VTT.
DownloaderAI uses whisper.cpp on your Windows CPU to transcribe a trusted video selected from Library. Media is not uploaded to a cloud transcription service.
Download for WindowsInstall resources only when needed
Choose a trusted Library input and output format. DownloaderAI downloads the whisper runtime and selected model on demand, then reuses successfully installed resources for later offline processing.
Workflow
How to use this workflow
Select a trusted Library video
Choose an eligible local media artifact from the ordinary Library workflow.
Choose a model and format
Select an available trusted model and TXT, SRT, or VTT output.
Install required resources
Allow the whisper runtime and model to download on demand if they are not already present.
Run local CPU transcription
Process the file on the Windows CPU and review the generated local output.
Use cases
Where it fits
Editable notes
Create TXT output for permitted research or production review.
Subtitle drafts
Create timestamped SRT output for later checking and editing.
Web captions
Create VTT output for an authorized media workflow.
Safety and privacy boundaries
- Media is processed locally and is not uploaded to a cloud transcription service.
- Inputs must come from the trusted ordinary Library workflow.
- Restricted content and Private Session media cannot enter Transcription in the current beta.
Known limitations
- Transcription is CPU-only; GPU acceleration is not promised.
- The whisper runtime and models are downloaded on demand and are not bundled in the installer.
- No specific speed, accuracy, minimum CPU, memory capacity, or instruction set is promised.
- Live installation of every model has not completed human verification, and Restricted Transcription is disabled.
Related DownloaderAI guides
Choose a video from Library
Understand trusted inputs, artifacts, and local file checks.
Why Private Session is separate
Private Session content does not enter ordinary Transcription records.
See the complete Windows workflow
Connect downloading, organization, and local transcription.
Download the Windows beta
Check the current version and unsigned-installer guidance.
Frequently asked questions
Does transcription upload my video?
No. DownloaderAI runs whisper.cpp locally on the Windows device.
Which output formats are available?
The current beta supports TXT, SRT, and VTT transcription output.
Does it use my GPU?
No. The current transcription workflow is CPU-only and does not promise GPU acceleration.
Are the model and runtime included?
No. The whisper runtime and selected model are downloaded on demand and can be reused offline after successful installation.
Can Restricted content be transcribed?
No. Restricted Transcription is currently disabled.
Download
DownloaderAI Desktop v1.3.0-beta.1 is available.
Visit the download page for the 64-bit Windows 10/11 beta installer and important SmartScreen information.