Speech-to-Text you can count on.

Don't settle for poorly supported APIs offered up by big tech. Start building with our high accuracy, state of the art Speech-to-Text API today.

Couple with Food

State of the Art Accuracy

Our API is powered by state of the art Deep Neural Networks. Our research team is constantly improving, and we release improvements every few weeks.

Customizable for Higher Accuracy

Boost accuracy for keywords and phrases, or share audio data with us for a custom-trained Acoustic Model.

Integrate in Minutes

Get started in minutes with our simple REST API using any language: Python, Node, Ruby, PHP, C#, etc.


License our Docker container to run the API on your own servers. Contact us for more info about license fees.

24x7 Customer Support

All customers get a dedicated Slack Channel with our engineers for 24x7 technical support and feedback.

Highly Scalable and Fast

Transcribe hundreds of audio files, or audio streams, in parallel with low latency.


We're supported by the top investors in Silicon Valley including Y Combinator and TechNexus Venture Collaborative.

Comes with all the core features you need.

High Accuracy

State of the Art accuracy on broadcast media, phone calls, and most types of audio.

Batch Transcription

Transcribe dozens of audio files in parallel. All file formats (WAV, MP3, M4A, etc.) are supported.

Real-Time and Synchronous Transcription

Transcribe speech in real-time over a WebSocket connection, or stream data to the Synchronous API for a transcript in the request-response cycle.


Boost accuracy for keywords and phrases, or share audio data with us for a custom-trained Acoustic Model.

Word Timings and Confidence Scores

See what was spoken when.

Automatic Punctuation and Casing

Make. Transcripts. More. Readable. Again.


Based on the latest Deep Learning research.

Our team is backed by Y Combinator's AI fund, and includes AI researchers from BMW, Cisco, and the open source community. We're constantly pushing out new neural networks based on the latest research in the AI community. In the future, we plan to publish some of our own papers.

Branch out from big tech.

Most companies rely on giant tech companies for AI services. It doesn't have to be this way. We don't surveil your users. We don't harvest your data. We won't suddenly deprecate our API. Together, we can make it so that AI technology doesn't end up in the hands of a few trillion dollar companies.

24x7 Support
Privacy Focused
Developer Friendly

Ready to start building?

Contact us or sign up now.

Trusted by over 7,000 developers and counting

AssemblyAI is hands-down the best voice recognition API we could find for our app.

Jakub  — Founder, VoiceStory

Ready to get started?

Contact us or sign up now.