Quickstart

Make a voice say a line, then transcribe it back, with nothing but an API key and a terminal.

1. Create an API key

Open API Key, choose Create API key and copy it: it is shown once. Keep it in an environment variable rather than in code.

bash

export YOOB_API_KEY="yoob_live_..."

2. Make speech

Pick a voice (the TTS page has samples of each) and send a line. Tags in square brackets set the feeling and add sounds.

bash

curl https://tts-api.yoob.com/v1/audio/speech \
  -H "Authorization: Bearer $YOOB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "yoob-tts-1", "voice": "w1243", "input": "[joy] Hey! [laughing] You actually came!", "response_format": "wav"}' \
  --output hello.wav

Play hello.wav. The first render after a quiet spell can take a second longer; after that, short lines come back in well under a second with "quality": "fast".

3. Transcribe it back

Send the file to speech to text. verbose_json returns the transcript with its emotion, word timings and a human-or-AI score: expect a low p_human here, since the voice is synthetic.

bash

curl "https://tts-api.yoob.com/v1/audio/transcriptions?response_format=verbose_json" \
  -H "Authorization: Bearer $YOOB_API_KEY" \
  -H "Content-Type: audio/wav" \
  --data-binary @hello.wav

The examples send the audio as the raw request body with its Content-Type. A multipart upload, the way the OpenAI SDKs send it, works too; see Speech to text.

Next steps

Text to speech

Voices, tags, quality and streaming.

Speech to text

Files, live audio, and everything a transcript carries.

Voice agents

Put listening and speaking together into a conversation.

Authentication

Keys, where they belong, and how usage is billed.