Highly Accurate Speech-to-Text

AssemblyAI is the most accurate Speech-to-Text API for transcribing
audio from phone calls, speeches, conferences, podcasts, and more.

Announcing Improved Accuracy - Read the Blog Post

Get up and running in minutes

Free for hackers, and no complicated sales calls or contracts for companies. Use the simple REST API to transcribe your audio. When you need help, our developers are in Slack 24/7 to provide support.

Train custom models for extremely high accuracy

In a single API call, create a custom model that'll be more accurate and tailored for your application. No machine learning or AI experience required. Custom models can also recognize as many application-specific words, like person or product names, as you need to.

Built for any voice application

From phone call analytics to voice assistants, and everything in between, use the API to put accurate speech recognition into your application. With timestamps, confidence scores, unlimited custom word support, and a simple API, you can quickly build cutting edge voice applications.