MAI Voice 2 API, MAI Transcribe 1.5 API: Pricing, Documentation

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/mai-voice-2/v1/text-to-speech' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"text": "Hello from Pixazo. This is MAI Voice 2.", "voice": "en-US-Harper:MAI-Voice-2", "style": "excited", "styledegree": 1}'

Callback Payload (success)

{
  "request_id": "microsoft-mai-voice-2_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "microsoft-mai-voice-2",
  "output": {
    "media_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp3",
    "media_type": "audio/mpeg"
  },
  "created_at": "2026-07-31T11:34:16.102Z",
  "completed_at": "2026-07-31T11:34:20.123Z"
}

Failure callback shape

{
  "request_id": "microsoft-mai-voice-2_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "microsoft-mai-voice-2",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on each status change.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Text limits

• text is required and must be a non-empty string of at most 8000 characters. Longer text is rejected at submit.
• Send plain text, not SSML — voice, style and styledegree are applied for you.
• Cost scales with the number of characters you submit (see Pricing).

Voices and Styles

47 prebuilt MAI-Voice-2 voices are available across 18 locales (de-DE, en-AU, en-US, es-ES, es-MX, fr-FR, hi-IN, hu-HU, it-IT, ko-KR, nl-NL, pt-BR, pt-PT, ro-RO, ru-RU, th-TH, tr-TR, zh-CN). Pass the full voice id in voice; every id ends in :MAI-Voice-2. The current list is maintained by Microsoft in the MAI-Voice documentation.

Styles

Supported styles are per voice. Most voices support the expressive set angry, confused, determined, disgusted, embarrassed, excited, fearful, happy, hopeful, jealous, joyful, regretful, relieved, sad, shouting, softvoice, surprised, whispering. Some voices instead support a narrative set (adventurous, caring, empathy, curious, encouraging, excited, friendly, cheerful, nostalgic, reflective, sad, disappointed, serious), and a few voices support no styles at all. Check the voice you are using against Microsoft’s per-voice style table before relying on a style.

Output

The completed request returns output.media_url, a 24 kHz mono MP3 (audio/mpeg). Fetch that URL to download the audio.

{
  "text": "Welcome to Microsoft Build. MAI Voice 2 supports multilingual expressive synthesis.",
  "voice": "en-US-Harper:MAI-Voice-2",
  "style": "happy",
  "styledegree": 1.2
}

Request Headers

Response Handling

A failure during synthesis is reported through the status endpoint as status: "ERROR", not as a synchronous error. Failed requests are not billed.

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED (or FAILED/ERROR), then fetch output.media_url for the audio.

curl 'https://gateway.pixazo.ai/v2/requests/status/microsoft-mai-voice-2_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "microsoft-mai-voice-2_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "microsoft-mai-voice-2",
  "output": {
    "media_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.mp3",
    "media_type": "audio/mpeg"
  },
  "created_at": "2026-07-31T11:34:16.102Z",
  "completed_at": "2026-07-31T11:34:20.123Z"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or FAILED/ERROR (failure).

Pricing

Billed at $0.022 per 1000 characters of input text, rounded up to the next 1000 characters. Billing is based on the text you submit, not on the length of the audio produced.

Because the character count is known when you submit, the cost of a request is determined at submit time. Failed requests are not billed.

Authentication

All requests require an API key passed via header.

Webhook (Optional)

Instead of polling, you can receive a Webhook callback when the request reaches a terminal state. Provide a Webhook URL via header on the submit request.

Example: enable Webhook

curl -X POST 'https://gateway.pixazo.ai/mai-transcribe/v1/speech-to-text' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook' \
  --data-raw '{"audio_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/audio/speech-17s.mp3", "locales": ["en"], "transcribe_style": "verbatim", "phrase_list": ["Acme Corp", "Dr. Yamada"]}'

Callback Payload (success)

{
  "request_id": "microsoft-mai-transcribe_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "microsoft-mai-transcribe",
  "output": {
    "media_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.json",
    "media_type": "application/json"
  },
  "created_at": "2026-07-31T11:34:16.102Z",
  "completed_at": "2026-07-31T11:34:20.123Z"
}

Failure callback shape

{
  "request_id": "microsoft-mai-transcribe_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "microsoft-mai-transcribe",
  "error": "Description of the failure"
}

Delivery semantics

• terminal mode: one Webhook callback when the request is COMPLETED or ERROR.
• sync mode: a Webhook callback on each status change.
• Callbacks are idempotent on request_id — de-duplicate on it.
• Respond 200 within a few seconds; the Webhook endpoint must be HTTPS.

Request Parameters

Audio limits

• The audio file must be under 50 MB and in WAV, MP3 or FLAC format.
• audio_url must be an http or https URL and must be publicly reachable.
• Speaker diarization is not supported — the transcript is not split by speaker.
• Prompt-tuning is not supported. Use phrase_list to bias recognition instead.
• Cost scales with the length of the audio (see Pricing).

Transcript Format

The completed request returns output.media_url, a JSON file containing the transcript. Fetch that URL to read the result.

{
  "text": "From Pixazo, this is MAI Voice 2 running through our own gateway.",
  "phrases": [
    {
      "offsetMilliseconds": 0,
      "durationMilliseconds": 5856,
      "text": "From Pixazo, this is MAI Voice 2 running through our own gateway.",
      "locale": "en",
      "confidence": 0
    }
  ],
  "duration_ms": 5860
}

Timestamps are returned per phrase, not per word: each entry in phrases carries offsetMilliseconds and durationMilliseconds.

Request Headers

Response Handling

An unreachable audio_url, an oversized file, or audio containing no detectable speech is reported through the status endpoint as status: "ERROR", not as a synchronous error. Failed requests are not billed.

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED (or FAILED/ERROR), then fetch output.media_url for the transcript.

curl 'https://gateway.pixazo.ai/v2/requests/status/microsoft-mai-transcribe_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "microsoft-mai-transcribe_019dxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "microsoft-mai-transcribe",
  "output": {
    "media_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/{request_id}/output.json",
    "media_type": "application/json"
  },
  "created_at": "2026-07-31T11:34:16.102Z",
  "completed_at": "2026-07-31T11:34:20.123Z"
}

Response Fields

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or FAILED/ERROR (failure).

Pricing

Billed at $0.006 per minute of audio ($0.36 per audio hour), rounded up to the next whole minute. Billing is based on the length of the audio you submit, not on the size of the transcript.

A hold is placed when the request is submitted, because the audio length is not known until transcription completes. The hold is reduced to the real cost once the duration is known, and released in full if the request fails.