hey everyone. I've been having an issue where sometimes the voice will have weird slurring, incorrect pronunctiation and other issues, anyone knows how to solve it? I'm using these settings and model. Any help is appreciated.

"model": "eleven_turbo_v2_5",
        "voiceId": "{{users_selected_voice}}",
        "provider": "11labs",
        "stability": 0.75,
        "similarityBoost": 0.75,
        "useSpeakerBoost": true,
        "style": 0.0,
        "optimizeStreamingLatency": 3
    },

It happens more consistently with specific words and names, but it can happen really anywhere and any word. This is an example of my assistant saying "goodbye" on a call, the line was meant to be "Have a great day. Goodbye." and it did this: https://cdn.discordapp.com/attachments/1498389238259912925/1498389239174402128/019dcfcb-8d2c-766b-bdf3-06717a9ecec5-1777307854164-f3cc3a0d-8c7d-4912-b748-6edb7ae46213-mono.wav?ex=69f0fb39&is=69efa9b9&hm=0ba251d2d6bc92855e4307754832505817d9d7adc1fe493ded9cfbc04368af59&

That sounds like classic TTS instability + phoneme misinterpretation, especially with streaming turned up. I’d start by lowering optimizeStreamingLatency (try 1 or 2) and slightly increasing stability (0.85–0.9) to reduce slurring. Also consider adding pronunciation hints or phonetic spelling for problem words/names. If it’s happening mid-sentence like “goodbye,” it could be chunking/streaming artifacts from VAPI rather than just ElevenLabs. I can help you debug this directly are you streaming partial text or sending full sentences at once? @alexdev

I'm using voice, so I'm pretty sure its streaming.

Got it, then it’s most likely streaming chunk issues, not just the voice model. Try buffering a bit more so you send full phrases, and lower optimizeStreamingLatency to avoid rushed output. Adding slight pauses or punctuation also helps fix words like “goodbye.” If you want, we can go over your setup privately I can help you pinpoint and fix it quickly. @alexdev

aight bet! I'll try that, thank you