Inconsistent Avatar Voice/Lip Sync & Action Text Appearing in Responses

Hi Convai team,

Issue 1 – Avatar voice/lip sync randomly not working

I build the project as a .exe and run it on my device. Many times the avatar does not speak at all, even though the response is correctly generated and displayed in the chat box.

When this happens:

  • The text response appears normally in the chat box

  • But the avatar does not:

    • play audio

    • trigger lip sync

    • change facial animations

    • play talking animations

The strange part is that the behavior is inconsistent:

  • Sometimes everything works perfectly

  • Other times the avatar stays completely silent

The same issue also happens when running the project directly inside the Unity Editor:

  • sometimes the avatar speaks normally

  • sometimes it only writes the response in the chat box without audio or animations

Issue 2 – Action/emotion text appearing in chat responses

Another issue with the newer version is that the avatar sometimes includes descriptive action text inside the response itself.

For example, the chat box shows something like:

adjusts posture with confident smile, then shifts slightly showing a more approachable demeanor

I would like the avatar to only speak normal dialogue responses and not include these action/emotion descriptions in the generated text.

Is there a recommended way to disable this behavior??

Any help would be greatly appreciated.

Thank you!!

Hi @K3
Thanks for the reply.

Is this for both issues that I have?

Yes, could you please share the session ID for the second issue?

I have the same issue in every session.

The specific one is session 96fed7fe4-5a66-4fb2-ab62-25b876f22013

Hi Convai team,

Kind reminder, if there is any update on the above.

Thank you!

This should already be fixed. Could you please test it again and let us know if you are still experiencing the issue?

Hi @K3

It works perfectly now, thank you very much.

Sorry for asking this maybe you don’t have the answer, but perhaps you can help point me in the right direction. Do you have any update regarding the issue where the avatar sometimes does not speak at all, even though the response is correctly generated and displayed in the chat box?

I’m asking because I’m planning to launch a project by the end of June and I’m trying to decide on the best direction technically. Are you planning to have this fixed before June or would you recommend proceeding with a WebGL based implementation instead?

Thank you again for the support.

Which LLM are you using?

Am using claude-4-5-sonnet (beta)

Did you try testing with other voice providers?

I’ve tried switching LLMs, but I don’t think I’ve tested other voice providers yet. Do you have any voice providers or LLM combinations you’d recommend trying? Are there any that you’d suggest avoiding?

We don’t have a specific recommended voice provider or LLM combination for this case.

We were not able to reproduce the issue on our side, so I’d recommend testing different voice providers and LLMs based on what best fits your character and use case.

If the issue occurs consistently with a specific combination, please share the details with us so we can review it further.

Hi Convai Team,

I am continuing to experience issues with my project’s consistency.

Most of the time, when I ask the avatar something, it responds in the chat box correctly, the talk animation is triggered, but the lip sync and voice do not work. Sometimes everything works except that the voice cannot be heard.

What makes this issue very strange is that it does not happen all the time, and I do not get any errors or warnings in the Unity Console. The problem occurs both when I run the project directly from Unity and when I build and run the .exe version.

I have also tested the Convai sample scene, and the exact same issue occurs there as well. In most cases, the lip sync and voice do not work, while all other features continue to function normally.

Could you please assist me in identifying the cause of this issue? If there are any logs, settings, or diagnostic steps you would like me to check. I have tried several voices and LLM’s.

Thank you!

Could you please share the session id?

Thank you for your response.

Below are some Session IDs where I experienced the issue. It happens very frequently and is actually more common for the voice and lip sync not to work correctly than for everything to function as expected.

7bbdb4ca-d8ec-4d16-b621-2333b86f2b36

c865becf-c512-4a3e-8609-bf1e10737093

e6aca68a-a038-447b-8d39-a2ca1045e354

I tested it again just now and encountered the issue once more. While this time the Unity Console has a warning (paste below). I’m not sure if it is related to the voice/lip-sync problem, but I wanted to share it in case it helps with the investigation.

The character with Unicode value \U0001F33F was not found in the [Inter_18pt-Medium SDF] font asset or any potential fallbacks. It was replaced by Unicode character \u25A1 in text object [Message].

This warning is related to the font, not directly to the voice or lip-sync issue.

The character being displayed is not available in the current TextMeshPro font asset or its fallback fonts, so Unity replaces it with a square symbol.

Please check the language/content being used and make sure your TextMeshPro font asset supports those characters, or add a suitable fallback font that includes them.

Thank you for the clarification.

As for the lip-sync problem, could you please let me know what I should check next?

Since this is the main issue I am trying to resolve.

I will inform you.