Speech to Text
Speech to Text turns what is said in an audio or video file into text using OpenAI’s transcription service. If it receives a video, it first extracts its sound. Long recordings are split into fragments, transcribed one after another, and the text is returned in full.
This is an older component. For new flows, use OpenAI Whisper, which also lets you set the language and transcribe several files at once.
When to use it
- Turn the recording of a meeting or a webinar into text to summarise it later.
- Transcribe the audio of a training video.
- Get the text of an interview to analyse it with an agent.
Configuration
You need an OpenAI API credential with your OpenAI account’s key. See Credentials.
| Field | What it does |
|---|---|
| Credential (required) | The OpenAI API credential. The audio is sent to OpenAI to be transcribed. |
| Input Audio o Video | The audio or video file you want to transcribe. Connect it to the component that brings the file. |
Output
The Output output returns the transcribed text, ready to pass to an agent, a summary or a component that stores it.
Example
A training academy records its lessons on video and wants a written summary of each one.
- Connect the lesson video to Input Audio o Video of Speech to Text.
- Choose the OpenAI credential.
- Connect Output to an agent that writes a summary with the main points of the lesson.
Good practices
- Every transcription uses your OpenAI account: the longer the recording, the more it uses.
- If you need to set the audio’s language or transcribe several files, use OpenAI Whisper.
Related components
- OpenAI Whisper: transcription with OpenAI, with language and several files.
- AssemblyAI: transcription with AssemblyAI.