llm
retrieve the LLM optimisation report
Analyses captured LLM traffic and reports optimisation signals (cost, latency, token usage, prompt and response characteristics) with an overall verdict. The report can also be emitted in evaluation-dataset formats for downstream tooling. An empty capture is a 200 with an empty report, not an error.
get/mockserver/llm/optimisationReport
Request
- Base URL: http://localhost:1080
- URL: http://localhost:1080/mockserver/llm/optimisationReport
- Auth: none declared
Query parameters
format'json' | 'markdown' | 'csv' | 'openai-evals' | 'fine-tune' | 'promptfoo'
output format; defaults to json. The dataset formats also accept the aliases evals (openai-evals) and finetune (fine-tune), and underscores in place of hyphens.
sessionstring
restrict the report to a single captured session
hoststring
restrict the report to a single upstream host
providerstring
restrict the report to a single LLM provider
Response
optimisation report returned in the requested format
object required
Changes
No recorded changes to this operation across all 1 revision of this API.