
by Google
Gemini Transcribe is Google's speech-to-text model in the Gemini API, with real-time streaming and pre-recorded audio transcription. It returns cleaned, formatted text with speaker labels and word timestamps.
APIContact for pricing
Step 1: Get a Gemini API key from Google AI Studio.
Step 2: Upload your audio file with the Files API.
Step 3: Send the file to the transcription model through the Interactions API.
Step 4: Set language codes, vocabulary hints or speaker options in the transcription config if needed.
Step 5: Read the output text, or use the Live API for streaming audio.
Yes. Real-time streaming runs through the Live API with sub-second latency.
Yes, for pre-recorded audio it attributes speech to speakers with timestamps. Support beyond three speakers is experimental.
It detects more than 85 languages automatically and handles switching between them.
Yes. You can pass up to 1,000 custom vocabulary phrases to bias recognition.
Maestra APIMaestra offers AI transcription, subtitling, translation, dubbing and live interpretation, with an API for working with files and exports in your own apps.