What provider to set for build.nvidia.com?

[BenceBakos]

They provide python snippet as well:

from openai import OpenAI

client = OpenAI(
  base_url = "https://integrate.api.nvidia.com/v1",
  api_key = "asd-123-asd"
)

completion = client.chat.completions.create(
  model="deepseek-ai/deepseek-r1",
  messages=[{"role":"user","content":""}],
  temperature=0.6,
  top_p=0.7,
  max_tokens=4096,
  stream=True
)

for chunk in completion:
  if chunk.choices[0].delta.content is not None:
    print(chunk.choices[0].delta.content, end="")

[taniashiba]

Hi @BenceBakos !

I think setting it up as an OpenAI-cpmatible provider is best here.

You can use environment variables:

export GOOSE_PROVIDER=openai
export OPENAI_API_KEY="your-nvidia-api-key"
export OPENAI_BASE_URL="https://integrate.api.nvidia.com/v1

Do you know where to get your API key for nvidia models? I think at build.nvidia.com under API you should be able to generate an API key to use in your config.

[BenceBakos]

With the following configuration:

GOOSE_PROVIDER: openai
GOOSE_MODEL: deepseek-ai/deepseek-r1
OPENAI_API_KEY: "nvapi-apikey"
OPENAI_BASE_URL: "https://integrate.api.nvidia.com/v1"

I get error:

Context: ○○○○○○○○○○ 0% (0/128000 tokens)
( O)> hellp
  2025-08-10T08:30:04.136981Z ERROR goose::session::storage: Failed to generate session description: Request failed
: invalid model ID (type: invalid_request_error) (status 400)
    at crates/goose/src/session/storage.rs:1302