What's new at AssemblyAI

The #1 Rated API for Automatic Speech Recognition

New
October 18, 2021

October 17th, 2021

New Async Model v8

  • Today, we're happy to announce the release of our most accurate Speech Recognition model to date—version 8 (v8)
  • This new model dramatically improves overall accuracy (up to 19% relative), and proper noun accuracy as well (up to 25% relative)
  • You can read more about our v8 model in our blog here. This new model dramatically improves accuracy around proper noun recognition and accent recognition without the need to specify an acoustic or language model.

Fixes

  • We fixed an edge case where a small percentage of short (<60 seconds in length) dual-channel audio files, with the same audio on each channel, were not collapsed to mono files, resulting in repeated words.
emoji negative reaction for 'October 17th, 2021' emoji neutral reaction for 'October 17th, 2021' emoji positive reaction for 'October 17th, 2021'
Thanks for your feedback
New
October 11, 2021

October 10th, 2021

New Real-Time v2 Model and Topic Detection v4 Model

  • We have released our Real-Time v2 model. This new model has improved the absolute accuracy of our real-time endpoint by ~10%.
  • Our new Topic Detection v4 Model also comes with significant accuracy improvements. With an accuracy boost of ~8.37%.

Check out our blog to read more about each model and its improvements!

emoji negative reaction for 'October 10th, 2021' emoji neutral reaction for 'October 10th, 2021' emoji positive reaction for 'October 10th, 2021'
Thanks for your feedback
New
October 03, 2021

October 3rd, 2021

New Topic Detection v3 Model, PII Redaction Fix

  • This week we have released our Topic Detection v3 model. 
  • This model improves on Topic Detection v2's ability to detect topics based on context. In the following text segment, the model was able to predict "Rugby" without the mention of the sport directly.
  • Instead of relying on the word "Rugby," the model was able to identify "Ed Robinson" as a Rugby coach and "six nations" as a Rugby tournament and correctly identify it as a conversation about Rugby.

Fixes and Improvements

  • We also released a fix for our PII Redaction feature that corrects an issue where the model would sometimes over-redact phone numbers as credit card information or social security numbers. 
  • Our model will now better identify phone numbers in cases where they are not explicitly referred to as a phone number—allowing them to be correctly redacted or unredacted based on the policies submitted with the POST request.
emoji negative reaction for 'October 3rd, 2021' emoji neutral reaction for 'October 3rd, 2021' emoji positive reaction for 'October 3rd, 2021'
Thanks for your feedback
New
September 26, 2021

September 26th, 2021

New Severity Scores for Content Safety

  • This week our team released an all-new feature for our Content Safety model! We now return a severity score along with the confidence and label keys in our response.
  • The severity score measures how intense a detected incident is on a scale of 0 to 1.
  • For example, a natural disaster that leads to mass casualties will have a score of 1.0, while a small storm that breaks a mailbox will only be 0.1.

Fixes and Improvements

emoji negative reaction for 'September 26th, 2021' emoji neutral reaction for 'September 26th, 2021' emoji positive reaction for 'September 26th, 2021'
Thanks for your feedback
Fix
September 19, 2021

September 19th, 2021

Misc. Fixes

This week, our engineering team has been focused on our v8 transcription model, which will introduce a major accuracy improvement across all audio types. Stay tuned! In the meantime, we shipped a few bug fixes around our real-time transcription API and the /v2/stream API.

  • Fixed an edge case where higher sample rates would occasionally trigger a "Client sent audio too fast" error from the real-time streaming API.
  • Fixed an edge case where some streams from real-time were held open after a customer idled their session. 
  • Fixed an edge case in the /v2/stream endpoint, where large periods of silence would occasionally cause automatic punctuation to fail.
  • Improved error handling for when a customer sends non-json input allowing us to communicate these occurrences more effectively. 

emoji negative reaction for 'September 19th, 2021' emoji neutral reaction for 'September 19th, 2021' emoji positive reaction for 'September 19th, 2021'
Thanks for your feedback
New
September 11, 2021

September 11th, 2021

Word Search Improvements, Punctuation added to /v2/stream

/v2/stream

  • You can now enable automatic punctuation when using the /v2/stream endpoint! This can be done by adding one extra parameter to your POST request (shown below).
  • For example:
  • {"audio_data: '"UklGRtjIAABXQVZFZ…", "punctuate": True}
  • Punctuation is disabled by default with the /v2/stream endpoint.
Word Search

  • Developers can now search for two-word phrases when using the new Word Search feature, such as "happy birthday" and "thank you."
emoji negative reaction for 'September 11th, 2021' emoji neutral reaction for 'September 11th, 2021' emoji positive reaction for 'September 11th, 2021'
Thanks for your feedback
We are ⚡by Beamer