StepFun: StepAudio 2.5 TTS (Text to Audio) API Reference

About

StepAudio 2.5 TTS is StepFun's contextual text-to-speech model: instructions (up to 200 characters) set the overall tone and persona, and text in parentheses inside the input gives inline directions that are not spoken. It supports Chinese, English, Japanese, Korean, French and Spanish (the last four in preview) and zero-shot cloned voices. Infron serves it through the OpenAI-compatible POST /v1/audio/speech endpoint with wav, mp3, flac, opus or pcm output and optional streaming. A request takes up to 1,000 characters. Billing counts a Chinese character as 1 and other non-space characters as 0.5.

1. Calling the API

Setup your API Key

To begin using Infron, create an account and generate your API Key.

export INFRON_KEY="YOUR_API_KEY"

Submit a request

Send JSON to POST https://llm.onerouter.pro/v1/audio/speech with your Infron API key. The request and response follow the OpenAI speech API, so the OpenAI SDKs work with the base URL https://llm.onerouter.pro/v1. The examples use the StepFun voice zixinnansheng. Replace YOUR_API_KEY with your key.

curl

curl https://llm.onerouter.pro/v1/audio/speech \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "stepfun/stepaudio-2.5-tts/text-to-audio",
    "voice": "zixinnansheng",
    "input": "Hello! Welcome to Infron.",
    "response_format": "mp3"
  }' \
  --output speech.mp3

Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    base_url="https://llm.onerouter.pro/v1",
    api_key="YOUR_API_KEY",
)

with client.audio.speech.with_streaming_response.create(
    model="stepfun/stepaudio-2.5-tts/text-to-audio",
    voice="zixinnansheng",
    input="Hello! Welcome to Infron.",
    response_format="mp3",
) as response:
    response.stream_to_file("speech.mp3")

TypeScript (OpenAI SDK)

import fs from "node:fs";
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://llm.onerouter.pro/v1",
  apiKey: "YOUR_API_KEY",
});

const response = await client.audio.speech.create({
  model: "stepfun/stepaudio-2.5-tts/text-to-audio",
  voice: "zixinnansheng",
  input: "Hello! Welcome to Infron.",
  response_format: "mp3",
});
fs.writeFileSync("speech.mp3", Buffer.from(await response.arrayBuffer()));

Response

A successful request returns HTTP 200 with the audio file as the body, not JSON. Content-Type follows response_format: audio/mpeg for mp3, audio/wav for wav, audio/flac for flac, audio/opus for opus, audio/pcm for pcm. The examples above save it as speech.mp3.

Streaming

Add "stream_format": "sse" to receive Server-Sent Events while the audio is generated. Each speech.audio.delta event carries the next base64 chunk of the audio in response_format; decode the chunks and append them in order. The stream ends with a speech.audio.done event that reports usage.

curl https://llm.onerouter.pro/v1/audio/speech \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "stepfun/stepaudio-2.5-tts/text-to-audio",
    "voice": "zixinnansheng",
    "input": "Hello! Welcome to Infron.",
    "response_format": "mp3",
    "stream_format": "sse"
  }' \
  --no-buffer
data: {"type":"speech.audio.delta","audio":"<base64 audio chunk>"}

data: {"type":"speech.audio.delta","audio":"<base64 audio chunk>"}

data: {"type":"speech.audio.done","usage":{"input_tokens":11,"output_tokens":0,"total_tokens":11}}

In this example input_tokens is the billable character count of the input.

Routing and usage

This model is served by stepfun (StepFun). Billing uses billable characters: a Chinese character counts as 1 and every two other non-space characters count as 1. The audio response carries no usage; with streaming, speech.audio.done reports the billable count as usage.input_tokens.

Errors

Parameters this model does not accept are rejected with HTTP 400 before the request is sent to StepFun, and nothing is charged. error.message explains the problem and ends with the request ID:

{
  "error": {
    "message": "voice is required (StepFun voice id) (request id: 20261111221810441285700pZnPkvCJ)",
    "type": "infron_ai_error",
    "param": "",
    "code": "invalid_request_error"
  }
}

Errors reported by StepFun, such as an unknown voice ID, are returned with StepFun's message. If a stream fails after it has started, an error event is sent instead of speech.audio.done.

2. Schema

Input

ParameterTypeRequiredDefaultRangeDescription
modelstringYes--stepfun/stepaudio-2.5-tts/text-to-audioModel ID.
inputstringYes--Up to 1,000 charactersText to synthesize. Text in parentheses, e.g. (lowered voice), is a non-spoken direction; do not wrap words that must be spoken in parentheses.
voicestringYes----StepFun voice ID, for example zixinnansheng. Official voices are listed in StepFun's TTS guide (see Sources); your own cloned voice IDs also work.
response_formatstringNomp3mp3, wav, flac, opus, pcmAudio format.
speednumberNo10.5 to 2.0Speaking rate.
instructionsstringNo--Up to 200 charactersOverall tone and persona in natural language. Sent as instruction.
languagestringNoenzh, en, ja, ko, fr, esLanguage of the speech; ja, ko, fr and es are in preview. Cloned voices support zh and en.
sample_rateintegerNo240008000, 16000, 22050, 24000, 48000Output sample rate.
stream_formatstringNoaudioaudio, ssesse streams Server-Sent Events (see Streaming).

StepFun fields

These fields are not part of the OpenAI speech API. Put them at the top level of the JSON body; with the OpenAI Python SDK pass them in extra_body.

ParameterTypeRequiredDefaultRangeDescription
volumenumberNo10.1 to 2.0Volume multiplier (0.1 = 10%, 2.0 = 200%).
pronunciation_maparrayNo----Pronunciation overrides: objects with a tone list of original/replacement entries, e.g. LOL/laugh out loudly; Chinese tones are numbered 1 to 5.
markdown_filterbooleanNo--true, falseFilter Markdown syntax out of the input.

Unsupported fields

FieldBehavior
bit_rate, text_normalizationIgnored.
Other fieldsNot forwarded to StepFun.

Output

ParameterTypeRangeDescription
Response bodybinarymp3, wav, flac, opus, pcmThe audio file in response_format (without stream_format=sse).
Content-Type headerstringaudio/mpeg, audio/wav, audio/flac, audio/opus, audio/pcmMIME type of the audio.
typestringspeech.audio.delta, speech.audio.done, errorStreaming event type.
audiostringBase64Next chunk of the audio (speech.audio.delta).
usage.input_tokensintegerNon-negativeBillable count of the input (speech.audio.done).
usage.output_tokensinteger0Always 0 for this model.
usage.total_tokensintegerNon-negativeSum of input and output.
errorobject--message, type and code of a failure (error event).

3. Detailed Pricing

ModelEndpointBilling unitPrice
StepFun: StepAudio 2.5 TTSstepfunPer 1M billable characters$85.00

Prices are in USD. Billing uses billable characters: a Chinese character counts as 1 and every two other non-space characters count as 1.

Sources: Create speech, TTS guide and voices, Model, Pricing.