Hi folks, I am using nova 3, general, en-US and during calls the transcript is very often wrong. When I use the deepgram endpoint directly with the audio of the call, the transcript is perfect. What is going on? I am having customers constantly complaining so please look into this.

Hi, could you please provide a call id where you experienced this issue? We would like to review the logs to identify the cause.

Hi Kyle, one example is 1006e24c-1ae5-48c0-915b-0569531647c3 the transcriber picked up "paper notarized" as "paper no rice". When I pass the audio snippet via deepgram directly I get the right translation with the "en" model and the wrong translation with the "multi" model as per screenshots https://cdn.discordapp.com/attachments/1394581613454823454/1394937562665058314/Screenshot_2025-07-16_at_10.02.18.png?ex=6878a057&is=68774ed7&hm=aec5625d9cc0f5130d6b33f64e14ec31605858dcbdd730ad909b8e6e179610e1& https://cdn.discordapp.com/attachments/1394581613454823454/1394937563000864968/Screenshot_2025-07-16_at_10.02.04.png?ex=6878a057&is=68774ed7&hm=1a2b10c42c67de1123354ced82e2417cc38f62aaf098c9c74fbaa8ea9283d085&

Hey Antonio Leet, I wanted to let you know that we're managing a high volume of support requests at the moment, so our response time might be a bit slower than usual. I truly appreciate your understanding and will get back to you as soon as possible!  Thanks again for your patience!