OpenAI-compat: openai-generic named client with /v1 base_url, not openai or openai-responses

Working config (provider "openai-generic", named client, base_url /v1)

First-party openai talks to api.openai.com. openai-responses posts /responses. The Chat Completions join is openai-generic. base_url is required on that provider and is the /v1 root (host appends /chat/completions). Origin-only 404s. The openai-generic/<model> shorthand skips the named client's base_url; point the function at the named client.

A Chat Completions host whose public catalog is GET https://api.pzero.studio/v1/models (no key) fits openai-generic if base_url is https://api.pzero.studio/v1.

client<llm> Pzero {
  provider "openai-generic"
  options {
    base_url "https://api.pzero.studio/v1"
    api_key "<Bearer key for that host>"
    model "deepseek-v4-flash"
  }
}

Keep embeddings off this host.