Pixazo API→Models→Openbmb VoxCPM2
Pixazo API→Models→Openbmb VoxCPM2

VoxCPM API - Text to Speech Voice Design API

by Openbmb

View as Markdown

VoxCPM API provides developers with powerful tools to integrate continuous, tokenizer-free speech synthesis into their applications. By operating in a continuous latent space rather than relying on discrete audio tokens, it enables deep contextual expressiveness, zero-shot voice cloning, and custom voice design without sacrificing audio fidelity or risking traditional quantization artifacts.

Get API Key
Openbmb VoxCPM2 API

Models Version

LIMITED TIME OFFER

Get $5 Free Credit on First Payment

No strings attached — add funds and get $5 bonus instantly

Claim Your $5 →

Openbmb VoxCPM 2.0 Text to Speech API Documentation

POST https://gateway.pixazo.ai/voxcpm/v1/text-to-speech

Authentication

All requests require an API key passed via header.

HeaderTypeRequiredDescription
Ocp-Apim-Subscription-KeystringYesYour API subscription key

Text to Speech - Openbmb VoxCPM2

Request Code

POST https://gateway.pixazo.ai/voxcpm/v1/text-to-speech
Content-Type: application/json
Cache-Control: no-cache
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

{
  "text": "Hello, from Pixazo.",
  "cfg_value": 2.0,
  "dit_steps": 10
}
import requests

url = "https://gateway.pixazo.ai/voxcpm/v1/text-to-speech"
headers = {
    "Content-Type": "application/json",
    "Cache-Control": "no-cache",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
}
data = {
    "text": "Hello, from Pixazo.",
    "cfg_value": 2.0,
    "dit_steps": 10
}

response = requests.post(url, json=data, headers=headers)
print(response.json())
const url = 'https://gateway.pixazo.ai/voxcpm/v1/text-to-speech';
const headers = {
  'Content-Type': 'application/json',
  'Cache-Control': 'no-cache',
  'Ocp-Apim-Subscription-Key': 'YOUR_SUBSCRIPTION_KEY'
};
const data = {
  text: 'Hello, from Pixazo.',
  cfg_value: 2.0,
  dit_steps: 10
};

fetch(url, {
  method: 'POST',
  headers: headers,
  body: JSON.stringify(data)
})
.then(response => response.json())
.then(data => console.log(data))
.catch(error => console.error('Error:', error));
curl -v -X POST "https://gateway.pixazo.ai/voxcpm/v1/text-to-speech" \
  -H "Content-Type: application/json" \
  -H "Cache-Control: no-cache" \
  -H "Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY" \
  --data-raw '{
    "text": "Hello, from Pixazo.",
    "cfg_value": 2.0,
    "dit_steps": 10
  }'

Output

{
  "output": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/openbmb-voxcpm2/1768578707564-851083.wav"
}
Try Now

Request Parameters - Text to Speech

Parameter Required Type Default Description
textYesstring—The text to convert to speech. Supports natural-language sentences and punctuation for prosody control.
cfg_valueNonumber2.0Classifier-free guidance scale. Higher values follow the text more strictly at the cost of naturalness.
dit_stepsNointeger10Number of diffusion-transformer inference steps. Higher values can improve quality but take longer.

Example Request

{
  "text": "Hello, from Pixazo.",
  "cfg_value": 2.0,
  "dit_steps": 10
}

Response

{
  "output": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/openbmb-voxcpm2/1768578707564-851083.wav"
}

Request Headers

Header Value
Content-Typeapplication/json
Cache-Controlno-cache
Ocp-Apim-Subscription-KeyYOUR_SUBSCRIPTION_KEY

Response Handling

Common status codes.

CodeMeaning
200Success — audio generated
Bad Request
401Unauthorized
402Insufficient Balance
403Forbidden
Too Many Requests
500Internal Server Error

Openbmb VoxCPM 2.0 Text to Speech API Pricing

ResolutionPrice (USD)
default$0

Openbmb VoxCPM 2.0 Text to Speech (Voice Design) API Documentation

POST https://gateway.pixazo.ai/voxcpm/v1/voice-cloning

Authentication

All requests require an API key passed via header.

HeaderTypeRequiredDescription
Ocp-Apim-Subscription-KeystringYesYour API subscription key

Voice Cloning - Openbmb VoxCPM2

Request Code

POST https://gateway.pixazo.ai/voxcpm/v1/voice-cloning
Content-Type: application/json
Cache-Control: no-cache
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

{
  "text": "Hello, this is a test.",
  "reference_audio_url": "https://your-audio-file.wav",
  "prompt_text": "transcript of reference audio (optional)"
}
import requests

url = "https://gateway.pixazo.ai/voxcpm/v1/voice-cloning"
headers = {
    "Content-Type": "application/json",
    "Cache-Control": "no-cache",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
}
data = {
    "text": "Hello, this is a test.",
    "reference_audio_url": "https://your-audio-file.wav",
    "prompt_text": "transcript of reference audio (optional)"
}

response = requests.post(url, json=data, headers=headers)
print(response.json())
const url = 'https://gateway.pixazo.ai/voxcpm/v1/voice-cloning';
const headers = {
  'Content-Type': 'application/json',
  'Cache-Control': 'no-cache',
  'Ocp-Apim-Subscription-Key': 'YOUR_SUBSCRIPTION_KEY'
};
const data = {
  text: 'Hello, this is a test.',
  reference_audio_url: 'https://your-audio-file.wav',
  prompt_text: 'transcript of reference audio (optional)'
};

fetch(url, {
  method: 'POST',
  headers: headers,
  body: JSON.stringify(data)
})
.then(response => response.json())
.then(data => console.log(data))
.catch(error => console.error('Error:', error));
curl -v -X POST "https://gateway.pixazo.ai/voxcpm/v1/voice-cloning" \
  -H "Content-Type: application/json" \
  -H "Cache-Control: no-cache" \
  -H "Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY" \
  --data-raw '{
    "text": "Hello, this is a test.",
    "reference_audio_url": "https://your-audio-file.wav",
    "prompt_text": "transcript of reference audio (optional)"
  }'

Output

{
  "url": "https://pub-xxx.r2.dev/voxcpm/abc123.wav",
  "audio_url": "https://pub-xxx.r2.dev/voxcpm/abc123.wav",
  "sample_rate": 48000,
  "elapsed_s": 2.3,
  "status": "done"
}
Try Now

Request Parameters - Voice Cloning

Parameter Required Type Default Description
textYesstring—The text to synthesize in the cloned voice. Supports natural-language sentences and punctuation for prosody control.
reference_audio_urlYesstring (URL)—Publicly accessible URL of the reference audio file (.wav recommended) that the model will clone the voice from. The reference should be a clean recording of a single speaker.
prompt_textNostring—Transcript of the reference_audio_url. Providing the exact transcript improves voice-cloning fidelity; if omitted, the model attempts internal transcription.

Example Request

{
  "text": "Hello, this is a test.",
  "reference_audio_url": "https://your-audio-file.wav",
  "prompt_text": "transcript of reference audio (optional)"
}

Response

{
  "url": "https://pub-xxx.r2.dev/voxcpm/abc123.wav",
  "audio_url": "https://pub-xxx.r2.dev/voxcpm/abc123.wav",
  "sample_rate": 48000,
  "elapsed_s": 2.3,
  "status": "done"
}

Response Fields

FieldTypeDescription
urlstringPublic URL of the generated .wav audio file (R2 CDN). Equivalent to audio_url.
audio_urlstringAlias of url kept for compatibility with audio-focused clients.
sample_rateintegerSample rate of the output audio, in Hz (e.g. 48000).
elapsed_snumberServer-side processing time in seconds.
statusstringAlways "done" on a 200 success. Error cases return non-2xx status codes (see Response Handling).

Request Headers

Header Value
Content-Typeapplication/json
Cache-Controlno-cache
Ocp-Apim-Subscription-KeyYOUR_SUBSCRIPTION_KEY

Response Handling

Common status codes.

CodeMeaning
200Success — cloned audio generated
Bad Request (missing/invalid text or reference_audio_url, unreachable reference URL)
401Unauthorized
402Insufficient Balance
403Forbidden
Too Many Requests
500Internal Server Error

Openbmb VoxCPM 2.0 Text to Speech (Voice Design) API Pricing

ResolutionPrice (USD)
default$0