Gather 2.0 Support #1323
Description
Ahoy!
You’re receiving this email from Twilio because you’ve used our existing speech capability more than 30 times in the past 3 months.
What do you need to know?
Twilio is upgrading its speech recognition system to improve accuracy, reliability, and flexibility for voice applications. To support this transition, we will begin migrating existing customers from 1.0 (the legacy speech recognition system) to the new 2.0 speech models starting April 1, 2025, through June 1, 2025. We expect a seamless migration with no interruptions to your service.
Here’s what this migration means for you:
New speech providers and models: Twilio is integrating Google’s latest V2 Speech APIs and Deepgram’s Nova2 models into .
Twilio picks default mode: If you do not choose a model or choose a generic speech model (i.e. one without provider name in it), the new system will select the best speech provider for each request based on parameters customers provide.
What do you need to do?
If you need more time: You can temporarily opt out of the migration by contacting Twilio Support before April 1, 2025. However, note that remaining on the legacy system will result in an increased pricing structure effective June 1, 2025; you can review the increased pricing structure on May 1, 2025. The opt-out period expires on September 30, 2025, after which all customers will be moved to the new system.
If you use Twilio SDKs: if your integration relies on 1.0, upgrade to the latest Twilio SDK before the migration to 2.0 to prevent compatibility issues upon recompiling.
For more details, check out our changelog, and the updated Twilio Docs on .
Sincerely,
Team Twilio
