Pagination

Every list answers one page, newest first (a run's pool, in ranking order), with a link to the next:

{
  "data": [ ... ],
  "links": { "next": "/api/v1/search?limit=20&cursor=WyIyMDI2LTEwLTAyVDE0OjAzOjExLjQwMloiLCI2ZjFjLi4uIl0" }
}

• links.next is a path with the same limit and any filter already on it. Request it as given, against the same base URL. When it is null, that was the last page.
• limit is 20 by default and at most 50 (50 and 100 for the audit trail). A number outside the range is clamped, and text is ignored.
• A cursor is opaque. Do not parse it, build one, or keep one across a long time: pass back exactly the one a page linked to. An edited or truncated one is 400 invalid-cursor.
• A cursor marks a place, not a snapshot. Rows added while you page appear on the first page and never shift a later one, so you will not see a row twice or skip one.

async function* eachSearch(headers: HeadersInit) {
  let next: string | null = '/api/v1/search?limit=50'
  while (next) {
    const page = await (await fetch(`https://liviamatch.com${next}`, { headers })).json()
    yield* page.data
    next = page.links.next
  }
}