Text by speech
What this tool does
Converts speech from an audio file into written text. Upload a link to an audio file, and the platform downloads it and uses an AI model to listen and type out what is said. You can choose the language or let the system guess it automatically. You can also specify if you want the exact words spoken or an English translation.
There is a 25 MB limit on audio file size. If the download fails or the file is too large, no text will be returned. The quality of the result depends on how clear the audio is and the model selected.
Common uses
- Transcribing meeting recordings or voice notes into text.
- Translating spoken content from one language into English.
Input data
| Parameter | Required | Example | Description |
|---|---|---|---|
url | Yes | https://remotion.media/audio.mp3 | The web address of the audio file. Must start with 'http' or 'https'. Example: 'https://example.com/audio.mp3'. This field is required. |
language | No | es | The language code for the speech, such as 'es' for Spanish. If left blank, the system tries to detect the language automatically. Example: 'en'. |
model | No | base | The power of the AI model to use. Smaller models are faster but less accurate. Options are 'tiny', 'base', 'small', 'medium', 'large-v2', or 'large-v3'. Example: 'base'. |
task | No | transcribe | Whether to transcribe the speech as spoken ('transcribe') or translate it into English ('translate'). Default is 'transcribe'. |
Output data
| Field | Description |
|---|---|
result | The transcribed or translated text. Example: 'Hello, how are you?'. If the audio is empty or fails, this will be empty. |