Feature Requests

For version v1 of the API would be great if you increased input text limit. Now it's 5,000 chars, 11labs offer 10,000-40,000, azure tts 20,000.

I was hoping to create custom voice using voice cloning using APIs. Any plans to support this in future ?

Should be able to set audio output device before the call starts and during a call

I'm currently fighting with the Hume React SDK because preventing audio clippings that are played back during assistant pause or muted audio is not possible with the SDK's current toolset. Even when I'm logging that is muted audio and is Assistant pause both are true, I'm getting little audio chunks played back if I caused interruptions multiple times in a row that bring the SDK out of sync. But this is a question of robustness.

I need a way to be able to track where/when session settings were applied and changed, and be able to diagnose changes in the EVI personality, etc.

As different users have different requirements for how much thinking time they might need between utterances and sentences, we should be able to set this value with the config. The documentation taught me to use pause and resume assistant, but using this to increase latency when the user is talking (and preventing premature responses from the assistant) causes many hard-to-handle side effects with the SDKs and the Hume API's behavior. It seems like this should be a native feature and not something each developer needs to develop on their own, because they need to assume and find out the exact behavior of SDK and API to make this aligned and synchronized.

HumeAudioChat.jsx:2377 Connection error details: {type: 'socket_error', message: 'We could not connect to the voice. Please try again.', error: undefined, timestamp: '2025-07-29T00:57:23.125Z', component: 'HumeAudioChat-main'} component : "HumeAudioChat-main" error : undefined message : "We could not connect to the voice. Please try again." timestamp : "2025-07-29T00:57:23.125Z" type : "socket_error" [[Prototype]] : Object It would be awesome if we could get reasons why th econnect failed, either api key, malformed parameters, etc.

I can see in the React SDK documentation that there's still an example showing auth being passed to VoiceProvider (2) : function App() { const apiKey = process.env.HUME_API_KEY; return ( <VoiceProvider auth={{ type: 'apiKey', value: apiKey }} configId={/ Optional: Your EVI Configuration ID /} > {/ ... /} </VoiceProvider> ) ; } Thank you for pointing out this discrepancy. If you'd like to report this documentation inconsistency, you can submit feedback to the Hume Feature Requests Board to help ensure the documentation stays up to date with the latest implementation patterns.

EVI easily picks up background noise and sounds (chatter, music, etc) which causes a lot of disruption to the conversation. Feature request for background noise filtering and speaker isolation. I tried doing so using the prompt but it didn't help. It doesn't pause the conversation, but the EVI is "listening" to the background and takes that as part of the conversation now