audio-cpp using a fish-audio doesn't work with referance audio/text

LocalAI version: localai/localai:master-gpu-hipblas sha256:f3ff922777a09ec7f63093df7482ebb15d3f94c0f8611d49143fe455b4f7144d

Environment, CPU architecture, OS, and Version: I'm doing this on unraid 7.3, using portainer. AMD Ryzen 9 5900X Radeon AI PRO R9700 Local AI is setup using distribution mode with nodes.

Describe the bug When I try and use TTS with a voice that was a referance audio with text using audio-cpp backend with a fish-audio family model I get this error Error: rpc error: code = Internal desc = Fish Audio prepare with inline reference audio requires reference_text option

To Reproduce My model yaml looks like this.

name: audio-cpp-chatterbox
backend: audio-cpp

flash_attention: true

known_usecases:
    - tts
    - audio_transform

tts:
    voice_cloning: true

parameters:
    model: audio-cpp/Fish-Audio-S2-Pro-GGUF/fish-audio-s2-pro-q8_0.gguf

the model is downloaded from https://huggingface.co/audio-cpp/audio.cpp-gguf/tree/main/Fish-Audio-S2-Pro-GGUF

Inside voice library I click "Use in Text to Speech" and then when I try and generate anything it will error Error: rpc error: code = Internal desc = Fish Audio prepare with inline reference audio requires reference_text option

Expected behavior to be able to use different models with audio-cpp like the fish-audio family of models

Logs node-backend-logs-audio-cpp-chatterbox-2026-08-20.json

Additional context There is another PR that is touching ref audio/text inside audio-cpp https://github.com/mudler/LocalAI/pull/11620 if changes are done from master then there might be merge conflicts.