Welcome to Salad Transcription
Unlock the full potential of your audio, video, and text content with Salad Transcription API and Transcription Lite. Whether you need high-accuracy, AI-enhanced transcription or a faster, budget-friendly alternative, we offer a solution tailored to your needs. Our APIs transform your media into accurate transcripts, translations, summaries, captions, and subtitles, enhancing accessibility and expanding your global reach. Get up to 12.5 free audio hours when you register. Sign up nowKey Features
- Wide Format Support: Compatible with common file formats like MP4, MOV, WAV, and MP3.
- Multilingual Transcription: Transcribe content in over 97 languages, including English, Spanish, Russian, Arabic, and more.
- Advanced AI Translation: Translate transcriptions into English and between multiple languages, including French, German, Italian, Portuguese, Hindi, Spanish, and Thai. (Only available in Salad Transcription API)
- Summarization: Automatically generate concise summaries of long transcripts for faster content analysis. (Only available in Salad Transcription API)
- Speaker Identification: Easily differentiate between multiple speakers for more precise transcripts.
- Multichannel Separation: Transcribe multichannel audio with speaker and channel identification. No extra cost, no number of channels limit, supports all languages.
- Time Coding: Generate sentence and word-level timestamps, crucial for accurate captioning and subtitling.
- SRT Output: Produce industry-standard SRT files ready for use in popular video editors and players.
- Seamless Integration: Integrate Salad Transcription API directly into your existing platform with flexible JSON instructions and customizable parameters.
Why Choose Salad Transcription API?
Whether you’re looking to:- Enhance Accessibility: Make your content accessible to all audiences, including those with hearing impairments.
- Expand Globally: Reach a worldwide audience with multilingual transcriptions and translations.
- Improve Engagement: Add captions and subtitles to boost viewer engagement and comprehension.
- Facilitate Analysis: Convert speech to text for easier indexing, searching, and analysis of audio content.
Accuracy
Salad Transcription API delivers industry-leading transcription accuracy, consistently outperforming many commercial providers. Across benchmarking tests in multiple languages, our models achieve over 90% average accuracy, including in non-English datasets—making us one of the most accurate transcription services on the market today. We use standardized metrics like Word Error Rate (WER) and benchmark against public datasets to ensure transparency and fairness in evaluation. View full accuracy benchmarks and Benchmarking Methodology.Choose the Right Transcription API
| Feature | Salad Transcription API | Transcription Lite |
|---|---|---|
| Transcription | ✅ High Accuracy | ✅ Faster, Cheaper |
| Multilingual Support | ✅ 97+ Languages | ✅ Core Languages |
| Translation to English | ✅ Yes | ✅ Yes |
| Speaker Identification | ✅ Yes | ✅ Yes |
| Time Coding | ✅ Yes | ✅ Yes |
| SRT Output | ✅ Yes | ✅ Yes |
| Summarization | ✅ Yes | ❌ Not Available |
| SRT Translation | ✅ Yes | ❌ Not Available |
| LLM Translation | ✅ Yes | ❌ Not Available |
| Insights (Summarization, Sentiment Analysis, etc.) | ✅ Yes | ❌ Not Available |
| Cost | Standard Pricing | Lower Pricing |
Turnaround Times
The Salad Transcription API operates as a batch, async service, not a real-time service. While most jobs are processed quickly, turnaround times can vary during peak usage when the queue holds thousands of requests.Salad Transcription API Processing Speed
Our multi-step processing model runs audio at approximately 5x the standard playback speed, enabling most transcriptions to be completed faster than the actual audio duration. We use a sequential long-form algorithm for transcription, prioritizing transcription accuracy over speed. We use a sequential long-form algorithm for transcription, prioritizing transcription accuracy over speed, as accuracy is our highest priority. During high-demand periods, there may be increased turnaround times, but our infrastructure is designed to auto-scale as needed to ensure no job exceeds a processing time of 2 hours.Standard Response Times for a 15-minute (20 MB) Audio Clip:
- ** < 5 minutes**: 95% of jobs
- ** 5-30 minutes**: 3% of jobs
- ** 30-90 minutes**: 2% of jobs
- ** > 90 minutes**: 1% of jobs
Transcription Lite Processing Speed
Transcription Lite is optimized for speed, processing audio at approximately 40x real-time speed, significantly reducing turnaround times for transcription-only jobs.Standard Response Times for a 15-minute (20 MB) Audio Clip:
- ** < 30 seconds**: 95% of jobs
- ** 30 seconds - 2 minutes**: 4% of jobs
- ** > 2 minutes**: 1% of jobs
Supported Languages
We offer comprehensive support for a wide range of languages through our API for transcription and translation to English. These include: Afrikaans, Arabic, Armenian, Azerbaijani, Belarusian, Bosnian, Bulgarian, Catalan, Chinese, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, Galician, German, Greek, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Kannada, Kazakh, Korean, Latvian, Lithuanian, Macedonian, Malay, Marathi, Maori, Nepali, Norwegian, Persian, Polish, Portuguese, Romanian, Russian, Serbian, Slovak, Slovenian, Spanish, Swahili, Swedish, Tagalog, Tamil, Thai, Turkish, Ukrainian, Urdu, Vietnamese, and Welsh. Although the model was trained on over 97 languages, we highlight those that consistently demonstrate high-quality results, with a word error rate (WER) of less than 50%, a common benchmark for speech recognition accuracy. For languages not listed, the model may still generate transcriptions, but the accuracy could fall below acceptable standards. For the full list of languages supported by Salad Transcription API, please refer to our guides page. Transcription Lite is optimized for core languages to provide faster processing with lower cost.Get Started Quickly
Ready to transform your media content? Get started in minutes:- Create a Free Account: Visit portal.salad.com to log in or create a free account. New users receive up to 12.5 free audio hours ($1) to transcribe at no cost!
- Learn How to Use Your Free Credits: Watch our quick tutorial to make the most of your free transcription hours.
- Integrate the API: Use our flexible instructions to integrate transcription services into your platform.
- Use the Python SDK: Quickly integrate Salad Transcription API with your Python applications using our official SDK. The SDK simplifies authentication, file uploads, job polling, webhook handling, letting you get started in just a few lines of code.
- Explore Examples: Use our API reference page for easy testing.
- Billing: Understand how charges are calculated for the transcription services you use. With the lowest rates on the market, Salad Transcription API provides cost-effective solutions without compromising on quality.
Experience the efficiency and accuracy of Salad Transcription API today, and elevate your content to new heights.