Control Exa content freshness with maxAgeHours and migrate off the deprecated livecrawl parameter

Documented steps

1. Decide the freshness requirement first: cache-only, cache-with-age-bound, or force a fresh crawl.
2. Use the integer 'maxAgeHours' field (range -1 to 720) on /contents, or inside the 'contents' object on /search.
3. maxAgeHours: 0 = always livecrawl the page fresh.
4. maxAgeHours: -1 = cache only, never livecrawl (fastest, cheapest, may return nothing for uncached pages).
5. maxAgeHours: N (positive) = serve cache if it is younger than N hours, otherwise livecrawl.
6. Omit maxAgeHours entirely = documented default behaviour, livecrawl only as a fallback when no cache exists. This is the recommended default.
7. Migrate legacy livecrawl values using the official mapping: 'always' -> maxAgeHours 0; 'never' -> maxAgeHours -1; 'fallback' -> omit the field; 'preferred' -> no exact equivalent, approximate with a low value such as maxAgeHours 1.
8. Tune 'livecrawlTimeout' (milliseconds, default 10000, maximum 90000) when crawling slow origins.
9. Docs: https://exa.ai/docs/reference/livecrawling-contents and https://exa.ai/docs/reference/get-contents