Feature Requests

When issuing an auth token for a user to chat with EVI2, it would be nice to specify a limit of how many seconds that auth token can chat. It would also be good to have a way of tagging these auth tokens and getting reliable reports of how many seconds of billing were associated with each auth token tag. The end goal is to be able to track and limit Hume billing in a reliable way that can be tracked and enforced for each of my users, even when those users are accessing the same chat configuration. The only thing different for each user is the Hume auth token and the language_model_api_key. In this situation the browser itself is not trusted and so reporting chat id or chat group id from the browser does not solve the problem.

Solution for building in video conferencing environment like google meet

Heygen, Tavus, etc

I am currently building a react app with @humeai/voice-react The problem is that everything is on the "client side" This makes it difficult to build secure apps. For example, I want that when the user clicks on the call button, a modal appears if they don't have enough credits. Since everything is on the "client side", I am not sure it is possible to build this feature securely. Another example, I want to decrease the credits by 1 for every minute in the call. Same problem, everything that gives me information about the length of the call can be manipulated on the client side. Is there a work around?

Dear HUME AI Team, Hello there! I am a loyal and enthusiastic user of HUME AI, and I must say, I am genuinely impressed by the real-time voice conversation functionality. It is intuitive, efficient, and, in many ways, even surpasses the advanced voice capabilities of other platforms, like ChatGPT. That being said, I have a few humble suggestions that I believe could further enhance the user experience: Background Conversation Support: It would be incredibly helpful if HUME AI could support real-time voice conversations even when the screen of my iPad is off. This would make it much more convenient for users to engage with the assistant hands-free. Improved Volume for Deep Questions: I’ve noticed that the voice volume for the “Deep Questions” feature is relatively low. Increasing the volume would significantly improve usability, especially in environments with ambient noise. Option for Non-Interruptible Conversations: I would love to see an option where users can choose to have uninterrupted conversations. While the current system allows for interruptions (which is great for flexibility), an option to disable interruptions would be fantastic for those who prefer a more seamless listening experience during certain interactions. Additionally, I’ve encountered a small bug that I wanted to bring to your attention: while having voice conversations with HUME AI, I’ve often noticed that the system stops outputting speech, even though it still shows "speaking" on the screen. This happens quite frequently, and I think resolving this issue would greatly improve the experience. Thank you for continuously delivering such an exceptional product. I’m confident that these suggestions, if implemented, will further elevate the experience for all users. If you'd like me to elaborate on any of the points mentioned above, I’d be happy to do so. I look forward to seeing how HUME AI continues to evolve and succeed. Warm regards, Jarrett

Is there a way to somehow access twilio caller number when using custom model in the configuration? I would like to recognize the caller based on phone number. Thanks

Thanks

Reference: https://discord.com/channels/991790422433210459/1194014543182053457/1316544556241522768

Currently there is only a prod api key - when doing a lot of testing, we find that these test logs overpopulate our Chat History (which is a very useful feature, btw). Ideally there would be a test API key (maybe with some amount of monthly free credits?) that would not populate the Chat History logs.