This message was deleted.

It seems likely that your issue might be related to the audio format or the way the audio is being streamed to LiveKit. Based on the LiveKit documentation, the

AudioSource

class in LiveKit expects audio frames to be in a specific format, typically with a sample rate and number of channels that match the expected configuration. If your server is crafting sound in MP3 format, you may need to convert it to a format compatible with LiveKit's audio processing, such as PCM with a specific sample rate and channel configuration. To create a custom TTS plugin for LiveKit, you should subclass the

TTS

class as suggested in the LiveKit Slack discussion. This involves implementing methods to handle the synthesis of text to audio frames that LiveKit can process. You can refer to the ElevenLabs plugin as a guide, ensuring that your plugin correctly handles audio data conversion and streaming. Would you like more detailed guidance on implementing a specific part of the plugin? Sources: - LiveKit documentation - LiveKit Slack discussion - ElevenLabs plugin 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.