How to Transcribe Audio to Text
A step-by-step tutorial for transcribing audio to text in Voice Creator Pro using uploads, YouTube URLs, and live recording.
Transcribe Audio to Text
In this tutorial you will use the STT (Speech to Text) tab to transcribe audio into text and export it as subtitles or timestamped JSON.
The whole process takes under a minute per audio source.
Prerequisites
- Voice Creator Pro, running either as the desktop app on your machine, or in your browser with VCP Cloud
- An audio file, a YouTube URL (desktop app only), or a microphone, depending on which workflow you follow
Workflow A: Transcribe an Uploaded File
Step 1: Open the STT Tab
Click Lab in the left sidebar, then select the STT tab at the top. You will see the Audio Source panel on the left, the Results panel on the right, and History below.
Step 2: Upload Your Audio
Click Upload in the Audio Source panel and select an audio file from your computer (WAV, MP3, or any common format).
Step 3: Set the Language and Speakers
Choose the correct language from the Language dropdown. If you are not sure or the audio contains multiple languages, leave it on Auto-detect.
If more than one person speaks, set Speakers to how many there are, or leave it on Auto. Each speaker gets a colored label in the transcript.
Step 4: Pick an ASR Model
Click the model name in the Results panel to switch between ASR models:
- Qwen3-ASR - Highest accuracy. Use the Quality switch to choose Lower Quality (faster) or Higher Quality.
- Parakeet v3 - Fast transcription across many languages
- Orukeet - Parakeet retrained for more accents and languages
On the desktop app, Parakeet v3 and Orukeet need an NVIDIA graphics card.
Step 5: Transcribe
Click Transcribe. The full transcription text appears in the Results panel once processing finishes.
Workflow B: Transcribe from a YouTube Video
YouTube import is available on the desktop app.
Step 1: Open the STT Tab
Click Lab in the left sidebar, then select the STT tab at the top.
Step 2: Paste the YouTube URL
Click YouTube in the Audio Source panel, paste the video URL, and let Voice Creator Pro extract the audio.
Step 3: Set Language and Model
Choose a language (or leave on Auto-detect) and select an ASR model by clicking the model name.
Step 4: Transcribe
Click Transcribe. The result appears in the Results panel. This is a quick way to pull text from interviews, podcasts, or any public video.
Workflow C: Record and Transcribe Live
Step 1: Open the STT Tab
Click Lab in the left sidebar, then select the STT tab at the top.
Step 2: Record from Your Microphone
Click Record in the Audio Source panel and speak into your microphone. Click Stop when you are done.
Step 3: Set Language and Model
Choose a language and ASR model as described above.
Step 4: Transcribe
Click Transcribe to see the text of what you just recorded.
Workflow D: Transcribe Audio Generated in Voice Creator Pro
Step 1: Open the STT Tab
Click Lab in the left sidebar, then select the STT tab at the top.
Step 2: Load from History
Scroll down to the History section below the panels. It lists the audio you generated in the Lab, such as TTS and Design results. Find the generation you want to transcribe and click Use to load it as the audio source.
Step 3: Transcribe
Set your language and ASR model, then click Transcribe. This is useful for verifying what was generated, creating subtitles for generated voiceovers, or repurposing generated audio into written content.
Fixing Speaker Labels
If the transcript gives a line to the wrong person, click that turn's speaker label and move the turn, or every turn by that speaker, to someone else. To fix a single word, right-click it. Pick an existing speaker or New speaker.
For recordings with lots of quick back-and-forth, open the Speaker detection menu and switch to Nemotron before you transcribe. On the desktop app, Nemotron needs an NVIDIA graphics card.
Exporting Results
Once you have a transcription, click Copy text to copy it, or export it in two formats:
- SRT - A standard subtitle file. Drop it into any video editor (Premiere Pro, DaVinci Resolve, CapCut, etc.) to add captions instantly.
- JSON - Includes word-level timestamps. Use this when you need precise alignment for programmatic workflows, custom subtitle styling, or audio editing tools.
Click the matching export button in the Results panel to download the file.
Tips
- Use History for previous generations. If you already created audio in the TTS or Design tab, click Use on it in the STT tab's History. No need to export and re-upload.
- Auto-detect is good, but explicit is better. If you know the language, select it manually. This gives the model a head start and can improve accuracy on short clips.
- Set the number of speakers. When you know how many people are talking, setting Speakers gives more accurate labels.
- Try a different ASR model if results are rough. Click the model name and switch. Different models handle accents, background noise, and speaking speeds differently.
- SRT for video, JSON for code. Pick SRT when you just need subtitles. Pick JSON when you plan to process timestamps programmatically.
Next Steps
- Speech to Text reference - Full details on every setting, audio source, and export format
- Clone a Voice - End-to-end tutorial for voice cloning
- Voice Cloning reference - Deep dive on cloning settings and models
Turn a Script into Multi-Voice Audio
Paste or upload a script with speaker labels and Voice Creator Pro detects each speaker automatically, so you can assign a voice to each one and generate multi-voice audio.
Text to Speech
Generate speech in the Lab's TTS tab with your saved voices, community voices, or built-in model voices.