Only create the EmulateUserStartedSpeakingFrame if we have received a transcription.

2025-07-14 17:38:03 -03:00
parent 8fd5576879
commit 727af2e6fb
2 changed files with 10 additions and 1 deletions
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -12,6 +12,11 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - For `LmntTTSService`, changed the default `model` to `blizzard`, LMNT's
  recommended model.

+### Fixed
+
+- Fixed an issue where, in some edge cases, the `EmulateUserStartedSpeakingFrame`
+  could be created even if we didn't have a transcription.
+
 ## [0.0.76] - 2025-07-11

 ### Added