Twilio incident

Issues using Speech-to-Text when using DeepGram for ConversationRelay

Major Resolved View vendor source →

Twilio experienced a major incident on August 3, 2026 affecting ConversationRelay US1 and Conversation Orchestrator, lasting 2h 43m. The incident has been resolved; the full update timeline is below.

Started
Aug 03, 2026, 04:47 PM UTC
Resolved
Aug 03, 2026, 07:30 PM UTC
Duration
2h 43m
Detected by Pingoru
Aug 03, 2026, 04:47 PM UTC

Affected components

ConversationRelay US1Conversation Orchestrator

Update timeline

  1. investigating Aug 03, 2026, 04:47 PM UTC

    We are currently investigating an issue regarding degradation to Conversation Relay. We will provide an update in an hour or as soon as more information becomes available.

  2. investigating Aug 03, 2026, 05:12 PM UTC

    We are continuing to investigate this issue.

  3. investigating Aug 03, 2026, 05:24 PM UTC

    We are continuing to investigate this issue.

  4. identified Aug 03, 2026, 05:39 PM UTC

    Engineering has identified the issue. Customers using Deepgram speech models nova-2 and nova-3 for Conversation Relay are hard down. Customers using Deepgram speech model flux or Google speech models are not impacted. Engineering teams are working closely with Deepgram to investigate and remediate the issue.

  5. identified Aug 03, 2026, 06:38 PM UTC

    We are continuing to work on a fix for this issue.

  6. monitoring Aug 03, 2026, 06:45 PM UTC

    We are observing recovery for customers using Deepgram speech models nova-2 and nova-3 for Conversation Relay. Engineering teams will continue to monitor closely with Deepgram to ensure recovery is sustained. We will provide another update in 2 hours or as sooner.

  7. resolved Aug 03, 2026, 07:30 PM UTC

    The issue with Deepgram speech models nova-2 and nova-3 for Conversation Relay has been resolved and it is operating normally at this time.