Construct a Grounded AI Advisor provider

Convenience factory that resolves the backend by 3-tier precedence (explicit type argument -> EVENTSTUDY_ADVISOR_PROVIDER env selector -> default "custom") and constructs the matching provider object. All three backends are wired: the in-process "custom" hook and the HTTP "openai" (OpenAI-compatible) and "anthropic" (native Anthropic Messages) providers.

Usage

provider(type = NULL, fn = NULL, model = NULL, base_url = NULL, ...)

Arguments

type: Provider type, one of "custom", "openai", "anthropic". When NULL (the default) it is resolved from the EVENTSTUDY_ADVISOR_PROVIDER env var, falling back to "custom".
fn: For type = "custom", the user function function(prompt, schema, ...) returning a list or character. Required for the custom branch.
model: Optional character model identifier (resolved from EVENTSTUDY_ADVISOR_MODEL when NULL).
base_url: Optional character base URL for HTTP providers (resolved from EVENTSTUDY_ADVISOR_BASE_URL when NULL).
...: Additional arguments forwarded to the provider constructor.

Value

A ProviderBase subclass instance.

Details

The "custom" branch needs no network and neither httr2 nor jsonlite. Keys are never read here; the HTTP providers resolve them at call time inside their own complete().

Examples

# Custom provider runs in-process, no network:
p <- provider("custom", fn = function(prompt, schema) "advice text")
p$complete("Summarise")$text
#> [1] "advice text"

if (FALSE) { # \dontrun{
# HTTP providers make live calls; never run in examples or on CRAN:
p <- provider("openai", model = "gpt-4o")
p$complete("Summarise these diagnostics")

a <- provider("anthropic", model = "claude-opus-4-8")
a$complete("Summarise these diagnostics")
} # }