TTS in five minutes

Generate your first Rime TTS audio clip in five minutes using cURL, Python, JavaScript, or TypeScript.

This guide shows you how to generate your first audio clip with Rime’s text-to-speech (TTS) API and experiment with different voices and speech customizations.

Prerequisites

To follow this guide, you need:

• A Rime API token: Create a free Rime account and copy your API key from the API Tokens page.
• A language runtime, depending on which tab you follow:
  • cURL — just a terminal (preinstalled on macOS and Linux).
  • Python — Python 3.10 or later.
  • JavaScript — Node.js 18 or later.
  • TypeScript — Node.js 18+ plus tsx (npm install -g tsx).

Code blocks in this guide are tabbed. Pick cURL, Python, JavaScript, or TypeScript in each block to follow your preferred language.

Create your script

If you’d rather paste a working file and read along, grab the full script below. Otherwise, continue step-by-step.

Full script

Create a file called rime_hello_world.py, rime_hello_world.js, or rime_hello_world.ts (or run the cURL version directly in your terminal) and paste the full script:

Full script (copy/paste)

cURL

curl --request POST \
  --url https://users.rime.ai/v1/rime-tts \
  --header 'Authorization: Bearer your_api_key_here' \
  --header 'Content-Type: application/json' \
  --header 'Accept: audio/wav' \
  --output output.wav \
  --data '{
    "text": "Hello! This is Rime speaking.",
    "speaker": "celeste",
    "modelId": "coda"
  }'

Step-by-step code

Create a file called rime_hello_world.py, rime_hello_world.js, or rime_hello_world.ts and import the required library modules:

Python

import json
import urllib.request

Next, we’ll create a request to the Rime API.

Set the request headers, specifying the Rime API key that you copied:

Python

RIME_API_KEY = "your_api_key_here"

headers = {
    "Accept": "audio/wav",
    "Authorization": f"Bearer {RIME_API_KEY}",
    "Content-Type": "application/json"
}

Configure a payload specifying the details of the request:

Python

payload = {
    "text": "Hello! This is Rime speaking.",
    "speaker": "celeste",
    "modelId": "coda"
}

This payload includes the three required parameters:

• text adds the text that the model converts to speech.
• speaker sets the voice that the agent uses (view your options on our Voices page).
• modelId specifies which model the agent uses. Use coda for our most realistic flagship voices, or mistv3 for fast synthesis with the classic Mist lineup.

The Coda API reference details the many optional parameters you can add to request payloads.

Now that you’ve created the headers and payload, make a POST request to the Rime API and write the streamed audio response to a file:

Python

data = json.dumps(payload).encode("utf-8")

request = urllib.request.Request(
    "https://users.rime.ai/v1/rime-tts",
    data=data,
    headers=headers,
    method="POST"
)

with urllib.request.urlopen(request) as response:
    with open("output.wav", "wb") as f:
        while chunk := response.read(4096):
            f.write(chunk)

print("Audio saved to output.wav")

Streaming allows audio playback to begin before generation completes. This enables conversational flow, as the user can start listening to responses before the entire audio clip has been generated.

Although streaming is less important in this example, because we’re writing to a file, it’s vital when using models in real-time conversations. If this sounds interesting, follow the LiveKit quickstart to create your first conversational agent.

Test the script

Run your script from the terminal:

Python

python rime_hello_world.py

On a successful run, the terminal displays a confirmation that your audio file has been saved:

'Audio saved to output.wav'

Choose a voice

Rime offers a range of voices with different personalities. To change the voice, update the speaker parameter in your request:

Python

payload = {
    "text": "Hello! This is Rime speaking.",
    "speaker": "orion",  # Try different voices here
    "modelId": "coda"
}

Browse all available voices on the Voices page.

Custom pronunciation

Custom pronunciation is supported on Mist v1 and v2 only. Coda, Arcana, and Mist v3 do not support phonemizeBetweenBrackets yet.

The mistv2 model lets you specify the pronunciation of brand names or uncommon words using Rime’s phonetic alphabet. Add the custom pronunciation in curly brackets and set phonemizeBetweenBrackets to true:

Python

payload = {
    "text": "Welcome to {r1Ym} labs.",
    "speaker": "peak",
    "modelId": "mistv2",
    "phonemizeBetweenBrackets": True
}

Use the Pronunciation tool on the dashboard to generate phonetic strings for any word.

Next steps

Now that you can generate TTS audio, try following the LiveKit quickstart guide to learn how you can set up a real-time conversation with an agent.

Check out these resources to get more familiar with Rime:

Models

Compare Coda (flagship) and Mist v3 (fast)

Voices

Browse all available voice options

Latency

Optimize for real-time performance

Coda Streaming API

Stream audio with our flagship model