| name | appnz-provider-nvidia |
| description | Use this skill when the user asks to use NVIDIA, provider nvidia, or any nvidia model through app.nz; covers model selection when present, BYOK setup, OpenAI-compatible routing, provider-specific endpoints, aliases, and debugging gateway headers. |
NVIDIA on app.nz
Route NVIDIA through app.nz with one app.nz key. Use this skill when a user asks for NVIDIA, nvidia, one of its model ids, aliases, product integrations, or BYOK setup for this provider.
Workflow
- Use
/api/gateway/models?provider=nvidia to confirm the current public model list if exact availability matters.
- Call app.nz at
https://app.nz/v1 with Authorization: Bearer pk_live_...; do not call https://integrate.api.nvidia.com directly unless the user explicitly asks to bypass app.nz.
- Store a user-owned upstream key with
POST /api/gateway/provider-keys using provider nvidia when BYOK is required. The upstream secret is never returned by app.nz.
- Prefer a model matching the task modality. If unsure, start with
nvidia/minimax-m2.7 or app/auto.
- Read
X-Gateway-Provider and X-Gateway-Model response headers when debugging routing.
BYOK
curl -sX POST https://app.nz/api/gateway/provider-keys \
-H "Authorization: Bearer pk_live_..." \
-H "Content-Type: application/json" \
-d '{"provider":"nvidia","api_key":"REDACTED_UPSTREAM_KEY"}'
Chat Example
curl -s https://app.nz/v1/chat/completions \
-H "Authorization: Bearer pk_live_..." \
-H "Content-Type: application/json" \
-d '{"model":"nvidia/minimax-m2.7","messages":[{"role":"user","content":"Give me one practical idea."}]}'
Models
| Model | Modality | Aliases | Caps |
|---|
| nvidia/deepseek-v3.2 | chat | deepseek-v3.2, deepseek-3.2 | stream, tools |
| nvidia/deepseek-v4-pro | chat | nvidia-deepseek-v4-pro, deepseek-v4-pro-free | stream, tools |
| nvidia/devstral-2-123b | chat | devstral-2, devstral-2-123b | stream, tools |
| nvidia/minimax-m2.7 | chat | minimax-m2.7, mm-m2.7 | stream, tools |
