Text to speech
Add narration, spoken responses and accessible audio to your content.
Try your own textSTARSG VOICE
Give your product a voice.Make every conversation count.
Build voice assistants, narrate content and serve multilingual audiences with speech recognition and text-to-speech APIs.
Take a moment by the Seine. Next, we’ll head into the lanes of Montmartre for coffee, art and a different view of Paris.
Listen, understand and respond.
Turn speech into text, process it in your application, then turn the response into audio.
Speech to text
Turn recordings and spoken input into text for summaries, captions and search.
ExploreA voice that fits your content.
Start with the language, then listen for the delivery that suits your audience.
Language
Check that the model supports your audience’s language and listen to a sample in that language.
Style
Compare clarity, pace and tone. A guide, a story and a spoken reply call for different deliveries.
Context
Try a passage from your own content before choosing a voice for longer narration or conversation.
A place for voice in every experience.
Voice assistants
Connect spoken questions, model responses and audio playback.
Video narration
Add a voiceover to product videos and short-form content.
Learning and listening
Make lessons, articles and longer reads available as audio.
Multilingual products
Choose speech models and voices for the languages your audience uses.
Put spoken information to work.
Turn recordings into text for meeting notes, captions and search, or pass the transcript to your next workflow step.
Meeting transcript example
Let’s confirm what we’re shipping in this release.
We’ll test the voice experience first and collect feedback.
I’ll put the results together and share the next steps.
Scripted transcript illustrating a meeting workflow; no audio is being recognized.
Add speech to the tools you already use.
Start with a small request, check the result, then connect it to your application.
- 01
Create an API key
Create your key in the console and keep it on your server.
- 02
Choose your model
Check supported inputs, parameters and pricing in the model documentation.
- 03
Make your first request
Try a representative input and review the output before putting it into production.
Questions, answered
When should I use speech recognition or text to speech?
Use speech recognition to turn audio into text, and text to speech to read text aloud. Combine them with a language model to build a voice assistant.
Which languages and audio formats are supported?
Support depends on the model. Check its documentation for languages, output formats and streaming options before integrating.