You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've downloaded the Linux binaries, and downloaded a model from the readme, but upon trying to generate a model from an image yields "Axios Error (400)"
Maybe I'm missing a setup step?
Steps to reproduce
Download latest appimage
Select a folder
Wait for setup to finish
Download model TripoSG
Create a new workflow with image input -> model -> add to scene
Run
Modly version
0.4.0
Operating system
Garuda Linux 7.0.11-zen1-1-zen
Screenshots & logs
AxiosError: Request failed with status code 400
The workflow used is simple:
The model:
Confirmation
I have searched existing issues and this has not been reported yet.
I can reproduce this on the latest version of Modly.
Thanks — this is an extension setup/registration issue, not an inference failure.
The AxiosError 400 fires immediately on submit, before any 3D work starts. It means the backend can't find the TripoSG generator in its registry — which happens when the extension's bundled environment didn't finish installing. Downloading the model (step 4) only fetches the weights; the extension's runtime (its embedded Python venv with torch / triposg / diso) is built separately when the extension is installed, and if that build failed the model registers as "errored" and the run is rejected with a generic 400.
Try this first: Models page → TripoSG → Repair (re-runs the extension install). Watch the logs during that step.
To help us nail the root cause, could you share:
The backend/terminal logs from clicking Repair (the [setup] … lines, especially any SSL/downl
Your GPU (NVIDIA / AMD / none) and whether you have an NVIDIA card at all.
One likely Linux culprit on our side: the embedded Python ships without a CA bundle, so the setup downloads (diso wheel + triposg source) can fail over HTTPS. We're also going to fix the misleading 400 — a failed install should say "extension needs Repair", not "unknown model".
Activity
cloudling8 commentedon Jun 25, 2026
yeah get the same on windows
Lorchie commentedon Jun 26, 2026
Thanks — this is an extension setup/registration issue, not an inference failure.
The AxiosError 400 fires immediately on submit, before any 3D work starts. It means the backend can't find the TripoSG generator in its registry — which happens when the extension's bundled environment didn't finish installing. Downloading the model (step 4) only fetches the weights; the extension's runtime (its embedded Python venv with torch / triposg / diso) is built separately when the extension is installed, and if that build failed the model registers as "errored" and the run is rejected with a generic 400.
Try this first: Models page → TripoSG → Repair (re-runs the extension install). Watch the logs during that step.
To help us nail the root cause, could you share:
One likely Linux culprit on our side: the embedded Python ships without a CA bundle, so the setup downloads (diso wheel + triposg source) can fail over HTTPS. We're also going to fix the misleading 400 — a failed install should say "extension needs Repair", not "unknown model".
iammojogo-sudo commentedon Sep 27, 2026
You likely, like Lorchie said, have to repair to rebuild the venv. Delete the pycache and venv and repair.
lightningpixel commentedon Nov 14, 2026
The extension's environment didn't finish installing. Click Repair on the extension to fix it. Reopen if it persists.