Thank you so much for using our widget and for leaving such thoughtful and constructive feedback!

Our primary goal with this widget is to offer a simple, accessible, and cost-effective way for streamers to incorporate AI companions into their streams. We wanted to avoid the hassle of requiring third-party software or subscriptions to various services, which can be complicated for users who don’t have advanced technical knowledge.

1. Robotic voices: You're absolutely right that the voices can sound a bit robotic. The voices currently come from StreamElements’ native TTS system, which is completely free for users and doesn't require any extra accounts (unlike services such as Eleven Labs). While we recognize they aren't as realistic, we opted for simplicity and zero cost to help make it accessible to everyone.

2. Previewing voices: This feature is actually already included! You can use the command (by default, '!say') to have the bot read any text aloud. You don't even need to be live to test it—you can send the command in your Twitch chat while offline and hear the voice in OBS. It doesn't use any OpenAI API tokens either, so no extra costs!

3. AI responding to microphone input: While we understand the appeal of this feature, it would require the companion to be a separate app installed on your computer. This would complicate the setup process and might feel risky for many users, which is why we’ve focused on keeping the widget browser-based and simple to use.

Thanks again for your feedback — it's insights like these that help us improve, and we're excited to keep building out more features for the community!