Skip to contents

Get the Last Prompt

Source: R/inspect.R
get_last_prompt.Rd

Returns detailed information about the most recent LLM call, including the prompt sent, response received, and metadata like tokens and cost. Works with module-based calls.

Usage

get_last_prompt()

Value

A dsprrr_prompt_inspection object containing:

  • prompt: The full prompt sent to the LLM

  • response: The LLM's response

  • model: The model used

  • tokens_in: Input tokens used

  • tokens_out: Output tokens generated

  • cost: Cost in USD (if available)

  • timestamp: When the call was made

  • source: The module class that originated the call

  • program_artifact_id: Exact executable program identity, when available

  • trace_context: Caller-supplied correlation context

Returns NULL if no LLM calls have been made.

Examples

if (FALSE) { # \dontrun{
# Make an LLM call
mod <- module(signature("question -> answer"))
run(mod, question = "What is 2+2?", .llm = ellmer::chat_openai())

# Inspect what happened
get_last_prompt()
#> --- Last Prompt -----------------------------------
#> System: Given the fields `question`, produce the fields `answer`.
#>
#> User: question: What is 2+2?
#>
#> --- Response --------------------------------------
#> Assistant: {"answer": "4"}
#>
#> --- Metadata --------------------------------------
#> Model: gpt-4o-mini | Tokens: 45 in, 12 out | Cost: $0.0001
} # }