Hi everyone,
I recently signed up for the annual Topaz Video AI subscription, and I’m hoping to tap into the community’s experience — because I’ve hit a wall that Topaz Support and I haven’t managed to solve together.
My first run went perfectly. I processed a 14-minute clip on Starlight Precise — it took about 24 hours, but the output plays back cleanly and the quality is great. No complaints there at all.
The trouble started when I tried the same clip on Proteus. The model downloads fine, but every single time it reaches the “Loading Model” stage, it throws an error and asks me to send logs to Support.
Here’s everything I’ve tried so far, following Support’s guidance:
- Sent the logs and a screenshot over to Topaz Support.
- On their advice, did a full uninstall with Revo Uninstaller and reinstalled from scratch — same error.
- They then suggested rolling back to an older build (1.6.0 or 1.5.0). I installed 1.6.0 and downloaded all 651 available models. The downloads completed with no problems, but Proteus still fails at exactly the same “Loading Model” step.
So after quite a few rounds back and forth, I’m a bit stuck — and, I’ll admit, a little frustrated.
I also ran my logs through Claude, which flagged this as a possible cause:
“On Windows 1.6.0, the /v2/app_state/unauthed update-check fires with an empty license_expiration_datetime and user_email before authentication completes, returns a 422, and intermittently breaks first-time local model loads.”
A couple of questions for the group:
- Is this a common issue? Has anyone run into the same “Loading Model” failure — especially with Proteus, when Starlight runs perfectly fine?
- If you’ve hit it and managed to get past it, what actually fixed it for you?
Any pointers at all would be hugely appreciated. Thanks so much in advance!
