Convai unreal engine: audio

Hi team,

I’m working on a MetaHuman project in Unreal Engine 5 on a Mac Studio (Apple M2 Ultra). A colleague set up the project before leaving, and I’m continuing the work without deep Unreal/Convai experience.

**What works:**

- MetaHuman is in the scene, positioned and facing the camera

- Natural idle pose (body animation is working)

- Convai Chatbot component added, Character ID connected

- ConvaiFaceSync component added, “Toggle Blendshape or Viseme” is enabled

- Text chat works: I type a message and the character responds with text

- Voice input works: I hold T, speak, and the character responds (in text)

**What doesn’t work:**

- No TTS audio output at all — the character never speaks aloud

- No lip sync / facial animation when responding

- Unreal itself can play audio fine (tested StarterContent sound files)

- Mac system audio works (YouTube, other apps)

- Microphone permissions for Unreal are granted in System Settings

**What doesn’t work:**

- No TTS audio output at all — the character never speaks aloud

- No lip sync / facial animation when responding

- Unreal itself can play audio fine (tested StarterContent sound files)

- Mac system audio works (YouTube, other apps)

- Microphone permissions for Unreal are granted in System Settings

**Character setup on convai.com:**

- Language: German

- Voice: assigned (tested multiple German / multilingual voices)

- Character was updated and saved

**Questions:**

1. Is there a known issue with Convai TTS audio on macOS (Apple Silicon / M2)?

2. Is there a setting in the Chatbot or FaceSync component that could be blocking audio playback?

3. Are there specific logs I should check to debug this?

Thanks in advance for any help!

Hello,

Welcome to the Convai Developer Forum!

Which version of the Convai Plugin are you using?

I’m using 3.6.0. thx

Could you please share the output logs?

the output log is too long to send it here (it says). suggestions?

Upload the file, and make sure to remove your API key from the logs.

AI_ASSIST_out logs.pdf (1.1 MB)

We have checked this on our side and confirmed that it is working properly on Windows. Could you please try again after changing the LLM and voice provider?

If the issue still persists, we kindly recommend switching to the Windows platform if possible.

Please also note that the Mac version has not been fully tested yet. As a result, Mac support is limited in the legacy version. As you may know, our V4 Beta plugin has already been released, and all upcoming updates will be delivered through V4.

At the moment, V4 does not support Mac. However, Mac support is planned for future versions.

Thank you for your understanding.