API · GPT Transcribe Speech to Text API: Pricing & Code

by OpenAI

Speech to Text

openai/gpt-transcribe/speech-to-text

184 languages
Realtime Transcription
Speech to Text

Call it with the model name openai/gpt-transcribe/speech-to-text on POST /jobs/createTask, from $0.00008.

Model nameopenai/gpt-transcribe/speech-to-text

EndpointPOST https://api.popnow.ai/api/v1/jobs/createTask

# 1. Create the task. Keep the Idempotency-Key if you retry after a network error.
export IDEMPOTENCY_KEY="$(uuidgen)"
curl -X POST https://api.popnow.ai/api/v1/jobs/createTask \
  -H "Authorization: Bearer $POPNOW_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $IDEMPOTENCY_KEY" \
  -d '{
  "model": "openai/gpt-transcribe/speech-to-text",
  "input": {
    "audio_url": "https://media.popnow.ai/shortfilm/voices/Aoede-46e4c427.mp3"
  }
}'

# 2. Copy data.taskId from the response, then poll until state is succeeded, failed, canceled or expired.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.popnow.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
  -H "Authorization: Bearer $POPNOW_API_KEY"

// Node 18+ (or any runtime with fetch). Keep the API key on your server.
const BASE = "https://api.popnow.ai/api/v1";
const headers = { Authorization: `Bearer ${process.env.POPNOW_API_KEY}` };
const sleep = (ms) => new Promise((r) => setTimeout(r, ms));

async function call(path, init = {}) {
  const res = await fetch(BASE + path, { ...init, headers: { ...headers, ...init.headers } });
  const body = await res.json();
  if (body.code !== 200) throw new Error(`${body.code} ${body.msg}`);
  return body.data;
}

const { taskId } = await call("/jobs/createTask", {
  method: "POST",
  headers: { "Content-Type": "application/json", "Idempotency-Key": crypto.randomUUID() },
  body: JSON.stringify({
    "model": "openai/gpt-transcribe/speech-to-text",
    "input": {
      "audio_url": "https://media.popnow.ai/shortfilm/voices/Aoede-46e4c427.mp3"
    }
  }),
});

// Poll with backoff. A finished task may still be copying files: wait while an asset is pending.
let task;
for (let delay = 2000; ; delay = Math.min(delay * 1.5, 15000)) {
  await sleep(delay);
  task = await call(`/jobs/recordInfo?taskId=${taskId}`);
  const copying = (task.output?.assets ?? []).some((a) => a.pending && !a.unavailable);
  if (["succeeded", "failed", "canceled", "expired"].includes(task.state) && !copying) break;
}
if (task.state !== "succeeded") throw new Error(`${task.state}: ${task.errorMessage ?? "no detail"}`);

// Each asset has a signed url (no API key needed) that expires at expiresAt: save the file.
for (const asset of task.output.assets) console.log(asset.url);

import os, time, uuid, requests

BASE = "https://api.popnow.ai/api/v1"
HEADERS = {"Authorization": f"Bearer {os.environ['POPNOW_API_KEY']}"}


def call(method, path, **kwargs):
    body = requests.request(method, BASE + path, headers={**HEADERS, **kwargs.pop("headers", {})}, timeout=30, **kwargs).json()
    if body["code"] != 200:
        raise RuntimeError(f'{body["code"]} {body["msg"]}')
    return body["data"]


task_id = call(
    "POST",
    "/jobs/createTask",
    headers={"Idempotency-Key": str(uuid.uuid4())},  # reuse it if you retry
    json={
        "model": "openai/gpt-transcribe/speech-to-text",
        "input": {
            "audio_url": "https://media.popnow.ai/shortfilm/voices/Aoede-46e4c427.mp3"
        }
    },
)["taskId"]

# Poll with backoff. A finished task may still be copying files: wait while an asset is pending.
delay = 2.0
while True:
    time.sleep(delay)
    task = call("GET", "/jobs/recordInfo", params={"taskId": task_id})
    assets = (task.get("output") or {}).get("assets") or []
    copying = any(a.get("pending") and not a.get("unavailable") for a in assets)
    if task["state"] in ("succeeded", "failed", "canceled", "expired") and not copying:
        break
    delay = min(delay * 1.5, 15.0)

if task["state"] != "succeeded":
    raise RuntimeError(f'{task["state"]}: {task.get("errorMessage", "no detail")}')

# Each asset has a signed url (no API key needed) that expires at expiresAt: save the file.
for asset in assets:
    print(asset["url"])

The input is built from the schema below, so it passes validation as written. Change the prompt and links to your own.

Price

Option	USD per second
Standard	$0.00008

Billed in USD · Every option and its price: GET /models/openai/gpt-transcribe/speech-to-text

Limits

Prompt length (characters): 2,048
Largest upload: 25 MB
Languages: 184

Parameters

These fields come from the same schema we check every request against. Fields that are not listed are rejected.

audio_urlstringRequired
The voice or music track to use.
Range
Character count: 1 – … · audio/mpeg, audio/wav, audio/mp4, audio/aac +3 · ≤ 25 MB
languagestringOptional
The language of the speech.
Allowed values
"aa""ab""ae""af""ak""am""an""ar""as""av""ay""az""ba""be""bg""bh""bi""bm""bn""bo""br""bs""ca""ce""ch""co""cr""cs""cu""cv""cy""da""de""dv""dz""ee""el""en""eo""es"+144
Range
Character count: 2 – 2
promptstringOptional
Describe what you want to make.
Range
Character count: 0 – 2048
language_hintsarray<object>Optional
Likely spoken language codes, one per row.
Range
Item count: 0 – 10
Fields inside
- language_hints[].codestringRequired
  Two-letter language code, for example en or fr.
  Range
  Character count: 2 – 2
keytermsarray<object>Optional
Names and terms that help the model follow your preferred spelling in the transcript.
Range
Item count: 0 – 100
Fields inside
- keyterms[].termstringRequired
  A word or short phrase to help with transcription.
  Range
  Character count: 1 – 100
temperaturenumberOptional
Variation between takes. Higher is less predictable.
Range
0 – 1

Raw JSON Schema

{
  "additionalProperties": false,
  "properties": {
    "audio_url": {
      "contentMediaType": [
        "audio/mpeg",
        "audio/wav",
        "audio/mp4",
        "audio/aac",
        "audio/ogg",
        "audio/opus",
        "audio/flac"
      ],
      "format": "uri",
      "minLength": 1,
      "type": "string"
    },
    "keyterms": {
      "description": "Names and technical vocabulary to recognize, one term per row.",
      "items": {
        "additionalProperties": false,
        "properties": {
          "term": {
            "description": "A name or technical term in its normal spelling, without angle brackets or line breaks.",
            "maxLength": 100,
            "minLength": 1,
            "pattern": "^[^<>\\r\\n]+$",
            "type": "string"
          }
        },
        "required": [
          "term"
        ],
        "type": "object"
      },
      "maxItems": 100,
      "type": "array"
    },
    "language": {
      "description": "Two-letter spoken language code.",
      "enum": [
        "aa",
        "ab",
        "ae",
        "af",
        "ak",
        "am",
        "an",
        "ar",
        "as",
        "av",
        "ay",
        "az",
        "ba",
        "be",
        "bg",
        "bh",
        "bi",
        "bm",
        "bn",
        "bo",
        "br",
        "bs",
        "ca",
        "ce",
        "ch",
        "co",
        "cr",
        "cs",
        "cu",
        "cv",
        "cy",
        "da",
        "de",
        "dv",
        "dz",
        "ee",
        "el",
        "en",
        "eo",
        "es",
        "et",
        "eu",
        "fa",
        "ff",
        "fi",
        "fj",
        "fo",
        "fr",
        "fy",
        "ga",
        "gd",
        "gl",
        "gn",
        "gu",
        "gv",
        "ha",
        "he",
        "hi",
        "ho",
        "hr",
        "ht",
        "hu",
        "hy",
        "hz",
        "ia",
        "id",
        "ie",
        "ig",
        "ii",
        "ik",
        "io",
        "is",
        "it",
        "iu",
        "ja",
        "jv",
        "ka",
        "kg",
        "ki",
        "kj",
        "kk",
        "kl",
        "km",
        "kn",
        "ko",
        "kr",
        "ks",
        "ku",
        "kv",
        "kw",
        "ky",
        "la",
        "lb",
        "lg",
        "li",
        "ln",
        "lo",
        "lt",
        "lu",
        "lv",
        "mg",
        "mh",
        "mi",
        "mk",
        "ml",
        "mn",
        "mr",
        "ms",
        "mt",
        "my",
        "na",
        "nb",
        "nd",
        "ne",
        "ng",
        "nl",
        "nn",
        "no",
        "nr",
        "nv",
        "ny",
        "oc",
        "oj",
        "om",
        "or",
        "os",
        "pa",
        "pi",
        "pl",
        "ps",
        "pt",
        "qu",
        "rm",
        "rn",
        "ro",
        "ru",
        "rw",
        "sa",
        "sc",
        "sd",
        "se",
        "sg",
        "si",
        "sk",
        "sl",
        "sm",
        "sn",
        "so",
        "sq",
        "sr",
        "ss",
        "st",
        "su",
        "sv",
        "sw",
        "ta",
        "te",
        "tg",
        "th",
        "ti",
        "tk",
        "tl",
        "tn",
        "to",
        "tr",
        "ts",
        "tt",
        "tw",
        "ty",
        "ug",
        "uk",
        "ur",
        "uz",
        "ve",
        "vi",
        "vo",
        "wa",
        "wo",
        "xh",
        "yi",
        "yo",
        "za",
        "zh",
        "zu"
      ],
      "maxLength": 2,
      "minLength": 2,
      "pattern": "^[a-z]{2}$",
      "type": "string"
    },
    "language_hints": {
      "description": "Likely spoken language codes, one per row.",
      "items": {
        "additionalProperties": false,
        "properties": {
          "code": {
            "description": "Two-letter language code, for example en or fr.",
            "maxLength": 2,
            "minLength": 2,
            "pattern": "^[a-z]{2}$",
            "type": "string"
          }
        },
        "required": [
          "code"
        ],
        "type": "object"
      },
      "maxItems": 10,
      "type": "array"
    },
    "prompt": {
      "description": "Spelling and brief context in the language of the recording.",
      "maxLength": 2048,
      "type": "string"
    },
    "temperature": {
      "description": "Sampling temperature for recognizing the recording.",
      "maximum": 1,
      "minimum": 0,
      "type": "number"
    }
  },
  "required": [
    "audio_url"
  ],
  "type": "object"
}

Responses, polling and callbacks

Every model works the same way: create a task, then poll it or wait for a callback.

Price it first

cURL

curl -X POST https://api.popnow.ai/api/v1/jobs/quote \
  -H "Authorization: Bearer $POPNOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "openai/gpt-transcribe/speech-to-text",
  "input": {
    "audio_url": "https://media.popnow.ai/shortfilm/voices/Aoede-46e4c427.mp3"
  }
}'

Send the same request body to POST /jobs/quote to see the estimated and maximum USD charge before running. A quote does not run the task or charge your balance.

Accepted

HTTP

HTTP/1.1 202 Accepted

{
  "code": 200,
  "msg": "success",
  "data": {
    "taskId": "task_01k3m8x9q2z4v7n5p6r8s0t1w3",
    "state": "queued"
  },
  "request_id": "req_example_create"
}

HTTP 202 means accepted: the maximum USD amount is held until settlement. When a task is accepted, we reserve the maximum quoted amount from your API balance. On completion, we charge for actual usage, never more than the hold, and release the remainder. If the task fails, the hold is released.

Check status

cURL

curl "https://api.popnow.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
  -H "Authorization: Bearer $POPNOW_API_KEY"

Poll every few seconds with backoff and stop once state is succeeded, failed, canceled or expired.

Finished task

JSON

{
  "code": 200,
  "msg": "success",
  "data": {
    "taskId": "task_01k3m8x9q2z4v7n5p6r8s0t1w3",
    "model": "openai/gpt-transcribe/speech-to-text",
    "state": "succeeded",
    "output": {
      "assets": [
        {
          "url": "https://cdn.popnow.ai/tasks/task_01k3m8x9q2z4v7n5p6r8s0t1w3/output.mp3?signature=FROM_RESPONSE",
          "expiresAt": "2026-10-17T12:20:00Z",
          "mime": "audio/mpeg",
          "durationSeconds": 30
        }
      ]
    },
    "settled": true
  },
  "request_id": "req_example_record"
}

Download from url before expiresAt. Need a fresh link later? Ask POST /common/download-url.

Verify a callback

Node.js

import { createHash, createHmac, timingSafeEqual } from "node:crypto";
import express from "express";

const app = express();
// express.raw keeps the exact bytes that were signed; express.json would not match.
app.post("/hooks/popnow", express.raw({ type: "application/json" }), (req, res) => {
  const timestamp = req.header("X-Webhook-Timestamp");
  const signature = req.header("X-Webhook-Signature") ?? "";
  const body = JSON.parse(req.body.toString("utf8"));
  const taskId = body.data?.taskId ?? body.task_id;
  const digest = createHash("sha256").update(req.body).digest("hex");
  const expected = createHmac("sha256", process.env.POPNOW_WEBHOOK_SECRET)
    .update(`${taskId}.${timestamp}.${digest}`)
    .digest("base64");
  const ok = expected.length === signature.length && timingSafeEqual(Buffer.from(expected), Buffer.from(signature));
  // Check the signature first, then reject anything older than five minutes.
  if (!ok || Math.abs(Date.now() / 1000 - Number(timestamp)) > 300) return res.status(401).end();
  res.status(200).end(); // answer fast, then fetch the task and store its files
});

With callBackUrl set, we post the finished task to you. Check the signature against the raw body before you trust it.

Send a local file

cURL

curl -X POST https://api.popnow.ai/api/v1/common/upload-url \
  -H "Authorization: Bearer $POPNOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"contentType": "image/png", "bytes": 482113}'

# PUT the file bytes to data.uploadUrl with every header in data.headers, then commit it:
curl -X POST https://api.popnow.ai/api/v1/files/$FILE_ID/commit \
  -H "Authorization: Bearer $POPNOW_API_KEY"
# The commit returns a popnow://f/... reference. Put it in any image, video or
# audio field of the task input instead of a public https URL.

Public https links work as they are. Local files go through upload-url and commit first.