Fetch clean page content for a list of known URLs with the Exa /contents API and handle per-URL failures

Documented steps

1. POST https://api.exa.ai/contents with 'x-api-key' header.
2. Body must supply exactly one of 'urls' (array, 1-100 items, each 1-2048 chars) or the older 'ids' array — the schema is a oneOf, so sending both is invalid.
3. Content options are TOP-LEVEL here, not nested: {"urls":[...], "text":{"maxCharacters":4000}, "highlights":{"query":"..."}, "summary":{"query":"..."}}. This differs from /search, where the same options live under a 'contents' object.
4. Optionally add 'extras': {links, imageLinks, richLinks, richImageLinks, codeBlocks} — each an integer 0-1000, default 0 — to pull out link lists or fenced code blocks.
5. Set 'subpages' (0-100) and 'subpageTarget' (string or array of strings, each 1-100 chars) to also crawl matching child pages such as 'pricing' or 'docs'.
6. Send the request and read results[] for content plus statuses[] for per-URL outcomes.
7. Iterate statuses[]: each entry has {id, status: success|error, source: cached|crawled, error?: {tag, httpStatusCode}}. Treat this as the real success signal.
8. Retry or drop individual URLs based on the error tag rather than failing the whole batch.
9. Docs: https://exa.ai/docs/reference/get-contents