How to use Vibe Transcribe
This guide walks through setup → first model → first transcript → export. If you are brand new, expect about ten minutes to get to a first result, depending on download speeds.
You will learn
- How to install on Windows, Mac, or Linux
- How to pick a starter Whisper model
- How to transcribe a file or a supported URL
- How to export in common formats
Video walkthrough
Step 1: Install Vibe
Start on the download page, grab the installer for your OS, then follow the platform tips there. If security prompts appear, read them carefully and only continue if you trust the source.
Step 2: Download a Whisper model
On first launch, Vibe needs a model file. Small is the usual starting point for a good mix of speed and accuracy. You can add more models later in Settings.
Step 3: Load your media
- Drag and drop a file into the window, or
- Use the folder icon to browse, or
- Use the URL flow when you want to pull audio from a supported site (internet required for that step).
Step 4: Check settings
- Language: pick the spoken language or auto-detect.
- Task: transcribe in the original language, or translate to English (Whisper’s built-in path).
- Speakers: enable if you want speaker labels and your recording suits it.
Step 5: Run transcription
Start transcription and let it finish. Speed depends on model size, file length, and hardware. If it feels slow, try a smaller model first or check performance settings.
| Example setup | Model | Rough feel |
|---|---|---|
| Modern laptop, CPU only | Small | Often a few times slower than real time |
| PC with supported GPU | Small | Often much faster |
Step 6: Export
When the transcript is ready, export to the format you need. SRT/VTT are common for captions; TXT/DOCX are common for notes and documents.
YouTube and other URLs
If your version supports URL import, paste a link and follow the on-screen flow. URL features need internet access to fetch media, then transcription can still run locally depending on your workflow.
Related
Download Vibe Transcribe — free
Open source · Works offline · No account needed · v3.0.19
Free downloadFrequently asked questions
How do I use Vibe Transcribe for the first time?
Download and install Vibe, then open the app and download a Whisper model (Small is recommended). Once downloaded, drag a file into the Vibe window or click the folder icon, then click Transcribe. Your transcript will appear in real time.
How long does Vibe take to transcribe?
Speed depends on your hardware and the model size. On a modern CPU with the Small model, Vibe typically transcribes at 2–4x real-time speed (a 10-minute audio takes 2.5–5 minutes). With GPU acceleration enabled, this can increase to 10–30x real-time speed.
Can I transcribe a YouTube video with Vibe?
Yes! Click the URL/link icon in Vibe, paste any YouTube URL, and click Transcribe. Vibe uses yt-dlp internally to download the audio, then transcribes it locally. The video audio is downloaded to your device and processed offline.
What file formats can Vibe Transcribe handle?
Vibe supports virtually all common audio and video formats: MP3, MP4, WAV, M4A, MKV, MOV, AVI, FLAC, OGG, OPUS, WEBM, and more — powered by FFmpeg under the hood.
How do I export my transcript from Vibe?
After transcription completes, click the Export button (download icon) to save as SRT (subtitles), VTT (web subtitles), TXT (plain text), HTML, PDF, JSON, or DOCX. SRT and VTT are ideal for video captions; TXT is best for notes and documents.
Can I translate audio to another language with Vibe?
Yes. In the transcription settings, change the Task from 'Transcribe' to 'Translate'. Whisper will then convert any input language directly into English text. For other target languages, you can use the AI summary feature with an Ollama or Claude API integration.
Does Vibe Transcribe support multiple speakers?
Yes. Enable Speaker Diarization in the settings panel before transcribing. Vibe will label different speakers as SPEAKER_00, SPEAKER_01, etc. This works best with clear audio and minimal background noise.
How do I speed up Vibe Transcription?
Enable GPU acceleration in Settings → Graphics Performance → High. Use the Small or Tiny model for faster (though less accurate) results. Close other resource-heavy applications while transcribing. On macOS with Apple Silicon, CoreML acceleration activates automatically.