Create transcription ​

POST https://openqiao.com/api/v1/audio/transcriptions

The Create transcription endpoint converts input audio into text (Speech-to-Text).

This endpoint takes a JSON request body. The audio is passed inline as base64-encoded bytes (or by URL) in the input_audio object. By default it returns a single JSON response with the transcript and token usage. Set stream to true to receive the transcript incrementally as Server-Sent Events instead.

Request headers ​

Authorization string * ​

Bearer Token authentication, formatted as Bearer $OPENQIAO_API_KEY.

Content-Type string * ​

Request content type, defaults to application/json.

Request ​

model string * ​

The model ID to use, in the format <provider>/<model-name>, for example qwen/qwen3-asr-flash.

input_audio object * ​

The audio to transcribe. Provide either data or url (at least one is required). Contains the following fields:

• data string Optional — the base64-encoded audio bytes. This is the raw base64 string only, not a data: URI.
• url string Optional — a publicly accessible URL to the audio file. Used when data is not provided.
• format string Optional — the audio container format. One of wav, mp3, flac, m4a, ogg, webm, aac. Defaults to wav when omitted.

language string Optional ​

The language of the input audio as an ISO-639-1 code (e.g. en, ja, zh). Auto-detected when omitted.

enable_itn boolean Optional ​

Whether to apply inverse text normalization (e.g. rendering spoken numbers as digits) to the transcript. Uses the model's default when omitted.

stream boolean Optional ​

Whether to stream the transcript, defaults to false.

• false — returns a single JSON response with the full transcript and token usage.
• true — returns the transcript as Server-Sent Events (Content-Type: text/event-stream), matching OpenAI's streaming format. The transcript text is delivered in transcript.text.delta events (each carrying an incremental delta), followed by a transcript.text.done event carrying the full text and token usage, and finally a [DONE] event.

Response ​

Non-streaming (default) ​

On success, the response is a JSON object with the transcript and token usage:

{
  "text": "OpenQiao is an LLM API aggregation service.",
  "model": "qwen/qwen3-asr-flash",
  "usage": {
    "input_tokens": 120,
    "output_tokens": 11,
    "total_tokens": 131,
    "seconds": 3
  }
}

• text — the transcribed text.
• model — the model slug used to produce the transcription.
• usage — token usage with input_tokens, output_tokens, and total_tokens. The optional seconds field reports the duration of the transcribed audio in seconds, when the provider returns it.

Streaming (stream: true) ​

The response is a text/event-stream of Server-Sent Events. Each event is a data: line containing a JSON object:

data: {"type":"transcript.text.delta","delta":"OpenQiao is "}

data: {"type":"transcript.text.delta","delta":"an LLM API aggregation service."}

data: {"type":"transcript.text.done","text":"OpenQiao is an LLM API aggregation service.","usage":{"input_tokens":120,"output_tokens":11,"total_tokens":131,"seconds":3}}

data: [DONE]

• transcript.text.delta — carries an incremental chunk of transcript text in delta. Concatenating all deltas yields the same text as the non-streaming text field.
• transcript.text.done — carries the full text and the token usage object with input_tokens, output_tokens, and total_tokens (and seconds when reported).
• [DONE] — marks the end of the stream.

Providers without a native incremental transcription API emit the full transcript as a single transcript.text.delta followed by transcript.text.done.

On error, the endpoint returns a JSON object instead:

{
  "error": {
    "code": "...",
    "type": "...",
    "message": "..."
  }
}

Errors use standard HTTP status codes (e.g. 400 for an invalid request such as a missing input_audio.data, 403, 404).

# Build the base64 audio from a local file, then POST it as JSON.
AUDIO_B64=$(base64 -w 0 speech.wav)

curl https://openqiao.com/api/v1/audio/transcriptions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENQIAO_API_KEY" \
  -d '{
    "model": "qwen/qwen3-asr-flash",
    "input_audio": {
      "data": "'"$AUDIO_B64"'",
      "format": "wav"
    },
    "language": "en"
  }'

import { readFile } from "node:fs/promises";

// Build the base64 audio from a local file.
const data = (await readFile("speech.wav")).toString("base64");

const response = await fetch("https://openqiao.com/api/v1/audio/transcriptions", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    Authorization: `Bearer ${process.env.OPENQIAO_API_KEY}`, 
  },
  body: JSON.stringify({
    model: "qwen/qwen3-asr-flash",
    input_audio: {
      data,
      format: "wav",
    },
    language: "en",
  }),
});

const { text } = await response.json();
console.log(text);

import base64
import os
import requests

# Build the base64 audio from a local file.
with open("speech.wav", "rb") as f:
    data = base64.b64encode(f.read()).decode("utf-8")

response = requests.post(
    "https://openqiao.com/api/v1/audio/transcriptions",
    headers={
        "Content-Type": "application/json",
        "Authorization": f"Bearer {os.environ['OPENQIAO_API_KEY']}",  
    },
    json={
        "model": "qwen/qwen3-asr-flash",
        "input_audio": {
            "data": data,
            "format": "wav",
        },
        "language": "en",
    },
)

print(response.json()["text"])

{
  "text": "OpenQiao is an LLM API aggregation service.",
  "model": "qwen/qwen3-asr-flash",
  "usage": {
    "input_tokens": 120,
    "output_tokens": 11,
    "total_tokens": 131,
    "seconds": 3
  }
}

To receive the transcript as Server-Sent Events, set stream to true:

curl https://openqiao.com/api/v1/audio/transcriptions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENQIAO_API_KEY" \
  -d '{
    "model": "qwen/qwen3-asr-flash",
    "input_audio": {
      "url": "https://example.com/speech.mp3",
      "format": "mp3"
    },
    "stream": true
  }'

Content-Type: text/event-stream

data: {"type":"transcript.text.delta","delta":"OpenQiao is "}

data: {"type":"transcript.text.delta","delta":"an LLM API aggregation service."}

data: {"type":"transcript.text.done","text":"OpenQiao is an LLM API aggregation service.","usage":{"input_tokens":120,"output_tokens":11,"total_tokens":131,"seconds":3}}

data: [DONE]