The main issue is that the react-native-voice package does not work with Expo out of the box. This package requires native code integration, but Expo’s workflow abstracts native code, making it impossible to integrate libraries that require native modules unless they are natively supported by Expo. While there is a workaround (ejecting your app to use this library), it comes at the cost of sacrificing Expo’s advantages.

If you have some experience with AI, you could also look at AI models like Whisper from OpenAI. If you do not have any experience with AI, you can try and use Google Cloud Speech-To-Text API, which I have implemented in one of my projects. It's relatively easy to integrate, but keep in mind that only the first 60 minutes per month are free. Beyond that, you'll be charged per minute of usage.