MetaHuman not animating or playing audio with Convai on Unreal Engine 5.7

Hello Convai Team,

I am currently working on a project using Unreal Engine 5.7 and integrating Convai with a MetaHuman character.

Here is my setup:

  • Unreal Engine version: 5.7

  • Convai Plugin: Installed and connected successfully

  • MetaHuman character imported correctly

  • Convai Character ID and Player ID are set properly

  • The Convai agent connects successfully and responses are received inside Unreal (text responses work correctly)

The issue:

  • When I ask a question, Convai responds correctly (connection works).

  • However, the MetaHuman does not play any talking animation.

  • There is no voice/audio output from the MetaHuman.

  • I have already assigned the correct talking animation / animation blueprint in the Convai character settings.

  • Lip sync and facial animation do not trigger.

So in summary:

  • :white_check_mark: Convai connection works

  • :white_check_mark: Responses are received

  • :cross_mark: No speech animation

  • :cross_mark: No audio output from MetaHuman

    and this is part of my log :(ConvaiSubsystemLog: Attendee ID: ConvAI-Bot, Data: {“data”:{“error”:“TTS generation error: 400 Currently, only Chirp 3: HD voices are supported for streaming synthesis.”,“fatal”:false},“label”:“rtvi-ai”,“type”:“error”}
    ConvaiSubsystemLog: Error: Error : ‘TTS generation error: 400 Currently, only Chirp 3: HD voices are supported for streaming synthesis.’.
    ConvaiSubsystemLog: Attendee ID: ConvAI-Bot, Data: {“data”:{“error”:“TTS generation error: 400 Currently, only Chirp 3: HD voices are supported for streaming synthesis.”,“fatal”:false},“label”:“rtvi-ai”,“type”:“error”}
    ConvaiSubsystemLog: Error: Error : ‘TTS generation error: 400 Currently, only Chirp 3: HD voices are supported for streaming synthesis.’.
    ConvaiPlayerLog: Finished talking)

Hello,

Welcome to the Convai Developer Forum!

Please switch your character’s LLM to Gemini 2.5 Live.
For the voice, choose one of the GCP voices whose name/description clearly specifies style and language, for example:

Despina (Smooth, Gentle Japanese Female Voice)

it’s not working

Please share your character ID.

d02f42bc-df20-11f0-a8ae-42010a7be027

Try ElevenLabs voices for Arabic. You’re currently using the beta plugin. Not every voice is supported.

I switched the character language to English, and now the audio and talking animations are working correctly. However, lip sync is still not functioning, and the MetaHuman does not look at the player/camera yet

Please refer to the documentation. As mentioned, this is a beta version, so some features are not supported.

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.