Hi Kaan,
Started the app this morning only to find that the Digital Human keeps repeating “I’ve reached my maximum thinking capacity, shall we go through this again.”
Using Claude 4.5 Opus as the Core AI.
Currently in transition between US and Europe.
Thanks,
Jeasy
Could you please share the logs?
Vivience20_20260530_035305_Default.log (24.2 KB)
I don’t have the UE project on my laptop, but this is the log from the build app. Also the LTM is still not working as expected.
The latency time is using Claude 4.5 Opus is over 5 sec+, is there any way this can be reduced? Which other Core AI model could I use with low latency but with equal emotional context as Claude Opus?
Unfortunately, there is no solution for this in the V3 plugin. The plugin has been completely rebuilt in V4 with a new infrastructure.
My recommendation is to upgrade to the V4 plugin, as it includes the latest improvements and will continue receiving updates.
Once in Norway, I’ll start rebuilding the project with the v4 plugin as recommended. The previous test was positive however the lip-sync mouth movement were weird and over too strong making the mouth deform as the digital human would speak.
Do you feel that issue has been corrected in the new release **Version 4.0.0-beta.21?
**
Thanks Kaan, for your support.