The voice-to-text feature is great, but when I enable it in Antigravity and start speaking, it sometimes takes a very long time for it to recognize my voice and start transcribing (it doesn’t start transcribing what I first said until 30-45 seconds later).
When it takes too long, there are usually spelling errors, but once it starts transcribing, if I keep talking, it continues to recognize and transcribe correctly.
It actually does a great job of converting speech to text, but when it starts recognizing too late, everything goes wrong.
In addition, when I’m in the middle of a task with Gemini and I start speaking to trigger the speech-to-text feature so I can prepare the next prompt, if Gemini asks me a question during that time, everything I’ve said gets canceled, and I end up having spoken for nothing. This can be really frustrating at times. Having to explain everything all over again is frustrating.