--llm-format-expects

Output expectations for the LLM.

Example

zshot -t llm_text --llm-format-model "anthropic" \
  --llm-format-expects "Yes,No" https://example.com

Details

JSON output: a JSON Schema the result must validate against. The schema is used as given — a $ref to a URL or file path does not resolve. Use $defs and inline subschemas.

Text output: a comma-separated set of acceptable answers, e.g. “Yes,No”; the output is

normalized to the matching answer.