#Model assistance
1 messages · Page 1 of 1 (latest)
Doable 👍
Start with version (just to avoid chasing old bugs):
openclaw --version
Then set Nemotron using whichever provider you’re using:
If you’re using NVIDIA directly
export NVIDIA_API_KEY="nvapi-..."
openclaw models list --provider nvidia --plain | rg -i nemotron
openclaw models set nvidia/<exact-model-id-from-list>
openclaw models status
If you’re using OpenRouter
export OPENROUTER_API_KEY="sk-or-..."
openclaw models list --provider openrouter --plain | rg -i nemotron
openclaw models set openrouter/<exact-provider>/<exact-model-id>
openclaw models status
Important: OpenRouter model refs include extra /, so keep the full ID exactly as listed.
If you get Model "... is not allowed", your allowlist is blocking it (agents.defaults.models), so add it there or clear the allowlist.
Docs: