Input_audio_buffer.speech_stopped events not firing reliably with VAD (Realtime API)

API
1

I’m using the GPT Realtime API with VAD enabled to test speech input → text output.

Until yesterday everything worked fine, but now I’m seeing a problem with the events:
• After sending audio, I consistently receive input_audio_buffer.speech_started.
• However, input_audio_buffer.speech_stopped is no longer firing reliably.
• In some rare cases it does fire, but in most cases no further events (including speech_stopped) are delivered after speech_started.

It looks like this started happening suddenly without any code changes on my side.

I’d like to understand what I should check or adjust in this situation:
• Are there any recent changes to VAD behavior or supported models that could affect speech_stopped?
• Is there any known issue where speech_started is emitted, but subsequent VAD-related events (speech_stopped, transcription events, etc.) are suppressed?
• Are there recommended settings (e.g. turn_detection type or parameters) to make sure speech_stopped is emitted reliably?
• Is there a way to debug why the server thinks speech has not stopped (e.g. extra logging, thresholds, silence duration)?

2

Hello @superplain did you ever get a solution to this ?

3

Hi @anindow and welcome to the developer community forum.

Usually this is due to excessive noise from your audio source, try recording your audio input and playing it back and see if you have a lot of background noise. you could also mute the mic with a push to talk style button, tat way your audio events will be clean and you can then see if you get the same issues.