bannerbannerExperience Now
WaveSpeed.ai
DashboardExploreInspiration
Home/Explore/OpenAI Models/wavespeed-ai/openai-whisper-turbo
speech-to-text

speech-to-text

OpenAI Whisper Large V3 Turbo | Multilingual Speech-To-Text API | WaveSpeedAI

wavespeed-ai/openai-whisper-turbo

Accurate speech-to-text with OpenAI Whisper Large v3 Turbo: multilingual transcripts with auto language detection and punctuation. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Hint: You can drag and drop a file or click to upload

If set to true, the function will wait for the result to be generated and uploaded before returning the response. It allows you to get the result directly in the response. This property is only available through the API.

Idle

{
  "text": "Welcome to WaveSpeed AI."
}

Your request will cost $0.0007 per run.

For $1 you can run this model approximately 1428 times.

ExamplesView all

README

OpenAI Whisper Turbo — Speech-to-Text

WaveSpeedAI’s Whisper deployment offers production-grade speech recognition built on OpenAI’s large-v3-turbo model. It transcribes audio into accurate text with multilingual support, noise robustness, and fast GPU inference.

⚡ Key Features

  • 50+ Languages Supported — including English, Chinese, Spanish, French, Arabic, Japanese, Korean, and more.
  • Automatic Language Detection — no need to specify the input language manually.
  • Context-Aware Transcription — understands sentence boundaries and speech flow naturally.
  • Accurate Punctuation & Capitalization — generates clean, readable text automatically.
  • Noise-Tolerant Recognition — performs well even in real-world, imperfect audio environments.

🎧 Supported Formats

  • Audio: MP3, WAV, M4A, FLAC
  • Maximum duration per file: Up to 1 hour recommended
  • Bitrate: ≥ 32 kbps for optimal accuracy

💰 Pricing

Just $0.0007 per second !!!

🚀 Quick Start

  1. Upload your audio (e.g., .mp3, .wav, .flac) or provide a direct HTTPS URL.
  2. Optionally specify language or leave as Auto for automatic detection.
  3. Add a prompt (optional) to guide the transcription style or context.
  4. Submit the request and get your transcription in seconds.

Example JSON Output:

{
  "outputs": {
    "text": "Hello everyone, welcome to the show."
  }
}

💡 Notes

  • For long-form transcription, split large audio into segments under 10 minutes for best performance.
  • The Auto language setting is recommended for multilingual datasets.
  • You can use prompts to adapt tone, style, or contextual vocabulary (e.g., medical, legal).
  • Whisper automatically handles noise, accents, and varied speech speed gracefully.