Name: StyleTTS 2 API: Text to Speech with Style Transfer
Brand: magicapi
SKU: api-magicapi-text-to-speech-api-adirik-styletts2
Availability: InStock
Author: API Market

Shashank Agarwal

# 🔊 Text-to-Speech API (adirik StyleTTS 2)

## Overview
StyleTTS 2 is a text-to-speech model that generates speech with style diffusion. This listing runs the version published by adirik.

Send the `text` to speak. Add a `reference` recording to copy its speaking style, and tune timbre (`alpha`), prosody (`beta`) and emotion (`embedding_scale`). Each request returns the URL of one MP3 file.

## Features
- **Text to speech**: one `text` field in, one MP3 file out.
- **Style from a reference**: pass a `reference` speech URL to copy its speaking style.
- **Timbre and prosody**: `alpha` (timbre) and `beta` (prosody), each 0 to 1, apply to long texts and reference speakers. Lower values follow the previous or reference speech more than the text.
- **Emotion**: higher `embedding_scale` (0 to 5, default 1) gives more pronounced emotion.
- **Quality and speed**: `diffusion_steps` from 0 to 50 (default 10).
- **Reproducible output**: a fixed `seed` (default 0) gives the same result for the same input.

## Use Cases
1. **Narration**: voice-overs for videos, tutorials and presentations.
2. **Spoken content**: audio versions of articles, notes or product descriptions.
3. **Voice prompts**: short spoken messages for apps, kiosks and phone menus.
4. **Consistent style**: match a reference speaker's delivery across many clips with `reference`.

## Visual Examples
### Example Input
A sentence about StyleTTS 2 itself, with `alpha` 0.3, `beta` 0.7, 10 diffusion steps, `embedding_scale` 1.5 and `seed` 0. No reference recording:

```json
{
  "beta": 0.7,
  "seed": 0,
  "text": "StyleTTS 2 is a text-to-speech model that leverages style diffusion and adversarial training with large speech language models to achieve human-level text-to-speech synthesis.",
  "alpha": 0.3,
  "diffusion_steps": 10,
  "embedding_scale": 1.5
}
```

### Example Output
The request returned one MP3 file with the sentence spoken by the default voice.

[Listen to the example output](https://blog.api.market/wp-content/uploads/2026/10/styletts2-output-b83knpd1-api-market.mp3)

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `weights` | string | No | none |  | URL of model weights fine-tuned on a new speaker. If provided, a reference speech must also be provided. If not provided, the default model is used. |
| `text` | string | Yes |  |  | Text to convert to speech |
| `reference` | string (URL) | No | none |  | Reference speech audio URL to copy the speaking style from |
| `alpha` | number | No | `0.3` | 0 to 1 | Only used for long text inputs or with a reference speaker. Sets the timbre of the speaker. Lower values sample the style from the previous or reference speech instead of the text. |
| `beta` | number | No | `0.7` | 0 to 1 | Only used for long text inputs or with a reference speaker. Sets the prosody of the speaker. Lower values sample the style from the previous or reference speech instead of the text. |
| `diffusion_steps` | integer | No | `10` | 0 to 50 | Number of diffusion steps |
| `embedding_scale` | number | No | `1` | 0 to 5 | Embedding scale. Higher values give more pronounced emotion. |
| `seed` | integer | No | `0` |  | Seed for reproducibility |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.001 per API unit.
- **1 API unit = 1 request with up to 4 seconds of processing time.** A request that finishes within 4 seconds costs one unit ($0.001).
- A request that runs longer bills one more unit for each started block of 4 seconds: more than 4 and up to 8 seconds is 2 units ($0.002), up to 12 seconds is 3 units ($0.003), and so on.
- **Example cost**: the example request took 2.3 seconds of processing, which is 1 unit = **$0.001**.
- Longer texts and more `diffusion_steps` take longer and can use more units.
- **Failed or cancelled predictions are not billed.**
- **GET requests to check prediction status** are free and don't count toward API unit usage.
- **No monthly subscription fee.**

## API Documentation

### Base URL
All API requests go to: `https://prod.api.market/api/v1/magicapi/text-to-speech-api-adirik-styletts2`

### Authentication
Send your api.market API key in the `x-api-market-key` header:
```
x-api-market-key: YOUR_API_MARKET_KEY
```

### Asynchronous Processing
**IMPORTANT**: This API is asynchronous. The flow is:
1. Send a POST request to `/predictions`. A successful request returns **201 Created** with a prediction `id` and a `status` of `starting` or `processing`.
2. Poll GET `/predictions/{id}` with that ID.
3. Keep polling while `status` is `starting` or `processing`.
4. Stop when `status` is `succeeded` (read `output`), `failed` or `canceled` (read `error`).

A polling interval of 1 second and an overall timeout of 600 seconds (10 minutes) work well.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/text-to-speech-api-adirik-styletts2/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "989cb5ea6d2401314eb30685740cb9f6fd1c9001b8940659b406f952837ab5ac",
    "input": {
      "beta": 0.7,
      "seed": 0,
      "text": "StyleTTS 2 is a text-to-speech model that leverages style diffusion and adversarial training with large speech language models to achieve human-level text-to-speech synthesis.",
      "alpha": 0.3,
      "diffusion_steps": 10,
      "embedding_scale": 1.5
    }
  }'
```

Response (201 Created):
```json
{
  "id": "b83knpd1mdrga0d1d0rtgs218r",
  "version": "989cb5ea6d2401314eb30685740cb9f6fd1c9001b8940659b406f952837ab5ac",
  "status": "starting",
  "created_at": "2026-10-23T09:57:25.822Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/text-to-speech-api-adirik-styletts2/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "b83knpd1mdrga0d1d0rtgs218r",
  "status": "succeeded",
  "output": "https://blog.api.market/wp-content/uploads/2026/10/styletts2-output-b83knpd1-api-market.mp3",
  "error": null,
  "created_at": "2026-10-23T09:57:25.822Z",
  "started_at": "2026-10-23T10:01:50.505Z",
  "completed_at": "2026-10-23T10:01:52.814Z"
}
```

`output` is a single URL string pointing to the MP3 audio file.

### Status Codes
- **201 Created**: prediction created (POST `/predictions`). This is the normal success response.
- **200 OK**: prediction details returned (GET `/predictions/{id}`).
- **400 Bad Request**: invalid input, for example a missing `text` or a `diffusion_steps` value above 50.
- **401 Unauthorized**: missing or invalid API key.
- **404 Not Found**: no prediction with that ID.
- **429 Too Many Requests**: rate limit exceeded.
- **500 Internal Server Error** or **503 Service Unavailable**: temporary server problem; retry later.

## Rate Limits
All plans allow 10 requests per second. Requests above that return `429 Too Many Requests`; back off before retrying and honour any `Retry-After` header. Polling a prediction with GET is free but still counts toward the per-second limit, so poll every 1 to 5 seconds rather than in a tight loop.

## Error Handling
- Check the HTTP status code of every request and read the response body for the error message.
- A prediction can also end with `status` set to `failed`; the `error` field then explains why. Failed predictions are not billed.
- Retry 429, 500 and 503 responses with exponential backoff. Fix the input before retrying a 400.

## Implementation Guide
### Example Code

#### JavaScript
```javascript
const apiKey = "YOUR_API_MARKET_KEY";
const baseUrl = "https://prod.api.market/api/v1/magicapi/text-to-speech-api-adirik-styletts2";

async function createAndWaitForPrediction(input) {
  const response = await fetch(`${baseUrl}/predictions`, {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "x-api-market-key": apiKey
    },
    body: JSON.stringify({ version: "989cb5ea6d2401314eb30685740cb9f6fd1c9001b8940659b406f952837ab5ac", input })
  });

  if (response.status !== 201) {
    throw new Error(`API error: ${response.status} ${await response.text()}`);
  }

  const prediction = await response.json();
  return pollForResults(prediction.id);
}

const delay = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

async function pollForResults(predictionId) {
  const maxTimeout = 600000;
  const pollInterval = 1000;
  const startTime = Date.now();

  while (Date.now() - startTime < maxTimeout) {
    const response = await fetch(`${baseUrl}/predictions/${predictionId}`, {
      headers: { "x-api-market-key": apiKey }
    });

    if (!response.ok) {
      throw new Error(`API error: ${response.status} ${await response.text()}`);
    }

    const prediction = await response.json();

    if (prediction.status === "succeeded") {
      return prediction.output;
    }

    if (prediction.status === "failed" || prediction.status === "canceled") {
      throw new Error(`Prediction ${prediction.status}: ${prediction.error || "Unknown error"}`);
    }

    await delay(pollInterval);
  }

  throw new Error("Prediction timed out after 10 minutes");
}

createAndWaitForPrediction({
  "beta": 0.7,
  "seed": 0,
  "text": "StyleTTS 2 is a text-to-speech model that leverages style diffusion and adversarial training with large speech language models to achieve human-level text-to-speech synthesis.",
  "alpha": 0.3,
  "diffusion_steps": 10,
  "embedding_scale": 1.5
}).then((output) => console.log(output));
```

#### Python
```python
import time
import requests

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/text-to-speech-api-adirik-styletts2"


def create_and_wait_for_prediction(input_data):
    headers = {"Content-Type": "application/json", "x-api-market-key": API_KEY}
    response = requests.post(
        f"{BASE_URL}/predictions",
        headers=headers,
        json={"version": "989cb5ea6d2401314eb30685740cb9f6fd1c9001b8940659b406f952837ab5ac", "input": input_data},
    )

    if response.status_code != 201:
        raise Exception(f"API error: {response.status_code} {response.text}")

    return poll_for_results(response.json()["id"])


def poll_for_results(prediction_id):
    max_timeout = 600
    poll_interval = 1
    headers = {"x-api-market-key": API_KEY}
    start_time = time.time()

    while time.time() - start_time < max_timeout:
        response = requests.get(f"{BASE_URL}/predictions/{prediction_id}", headers=headers)

        if response.status_code != 200:
            raise Exception(f"API error: {response.status_code} {response.text}")

        prediction = response.json()

        if prediction["status"] == "succeeded":
            return prediction["output"]

        if prediction["status"] in ("failed", "canceled"):
            raise Exception(f"Prediction {prediction['status']}: {prediction.get('error') or 'Unknown error'}")

        time.sleep(poll_interval)

    raise Exception("Prediction timed out after 10 minutes")


output = create_and_wait_for_prediction({
    "beta": 0.7,
    "seed": 0,
    "text": "StyleTTS 2 is a text-to-speech model that leverages style diffusion and adversarial training with large speech language models to achieve human-level text-to-speech synthesis.",
    "alpha": 0.3,
    "diffusion_steps": 10,
    "embedding_scale": 1.5
})
print(output)
```

## FAQ
- **Can I copy the style of a specific speaker?**
  - Pass a recording of that speaker as `reference`. Lower `alpha` and `beta` make the result follow the reference more closely.
- **What does `weights` do?**
  - It loads model weights fine-tuned on a new speaker. You must also send a `reference` recording when you use it. Leave it empty to use the default model.
- **How do I make the speech more expressive?**
  - Raise `embedding_scale` (default 1, up to 5) for more pronounced emotion.
- **What audio format is returned?**
  - One MP3 file, as a URL in `output`.
- **What did the example request cost?**
  - 1 API unit, $0.001, because it took 2.3 seconds of processing.
- **Is the 201 response from POST an error?**
  - No. 201 Created means the prediction was accepted. Use the returned `id` to poll GET `/predictions/{id}` until `status` is `succeeded`, `failed` or `canceled`.

## Contact Us
If you have any questions, reach out to us through the chat on the bottom right of the page. You can also email our support team at hello@magicapi.com.


Convert text to speech with StyleTTS 2, with optional style transfer from a reference recording and controls for timbre, prosody and emotion.