Pagination ​

The three helpers use different traversal styles. Choose one based on how you want to consume the data.

paginate ​

const result = await aniLink.anilist.paginate(
    (page, perPage) => aniLink.anilist.query.page.medias({ page, perPage, type: "ANIME" }),
    "media",
    { perPage: 50, maxPages: 10, concurrency: 4 }
);
console.log(result.items.length, result.pageCount, result.truncated);

paginatePages ​

for await (const page of aniLink.anilist.paginatePages((page, perPage) =>
    aniLink.anilist.query.page.medias({ page, perPage, type: "ANIME" })
)) {
    console.log(page.pageInfo.currentPage, page.media.length);
    if (page.media[0]?.id === 1) break; // early exit without buffering everything
}

paginateChunks ​

const chunked = await aniLink.anilist.paginateChunks(
    (chunk, perChunk) =>
        aniLink.anilist.query.mediaListCollection({
            userId: 542244,
            type: "ANIME",
            chunk,
            perChunk,
        }),
    "lists",
    { perChunk: 500, maxChunks: 20 }
);
console.log(chunked.items.length, chunked.chunkCount, chunked.truncated);

Options and clamps ​

Ordering and truncation guarantees ​

• Results always follow page or chunk order, regardless of completion order. Concurrency never changes that order.
• Scheduling stops only after you consume the terminal result in order, not when the request settles. A fetched page that reports hasNextPage: false (or a chunk that reports hasNextChunk: false) halts further scheduling once you have consumed it. The helper waits for already-launched requests to finish and discards their payloads.
• Page look-ahead never launches a page beyond the smallest pageInfo.lastPage a received page has reported. The helper would otherwise wait for those requests and discard their payloads, so skipping them saves rate-limit budget. The lastPage bound does not affect chunk traversals. MediaListCollection chunks carry no pageInfo.
• truncated is true when the traversal stops at maxPages or maxChunks before the source runs out. It is also true when the lastPage bound ends a page traversal whose last fetched page reports hasNextPage: true. For paginatePages, the last yielded page's pageInfo reports the same condition: hasNextPage: true at currentPage === lastPage. A generator has no result flag.
• paginateChunks continues while a fetched chunk reports hasNextChunk: true, up to maxChunks.

Cancelling look-ahead ​

Pass a signal to cancel the traversal and immediately abort in-flight look-ahead requests. If you break out of a paginatePages loop without passing a signal, the generator's finally block aborts those requests. You spend no rate-limit budget on their discarded payloads:

const controller = new AbortController();

for await (const page of aniLink.anilist.paginatePages(
    (page, perPage, signal) =>
        aniLink.anilist.query.page.medias({ page, perPage, type: "ANIME" }, { signal }),
    { signal: controller.signal, concurrency: 4 }
)) {
    if (page.media[0]?.id === 1) {
        controller.abort(); // cancel in-flight look-ahead, then break
        break;
    }
}

Custom Page documents ​

customPage runs the same engine over a caller-authored Page document, for collections whose field combination the generated page operations do not expose. See the custom queries guide for the document contract.

Incremental consumption ​

The eager variants (paginate, paginateChunks) collect every response before returning, so onPage and onChunk fire after the last response arrives. They do not reduce peak memory or release collected items incrementally. To process pages as they arrive or stop early, use paginatePages:

for await (const page of aniLink.anilist.paginatePages((page, perPage) =>
    aniLink.anilist.query.page.medias({ page, perPage, type: "ANIME" })
)) {
    for (const item of page.media) store.upsert(item);
}

Use onPage and onChunk to log or record metrics for each page or chunk after traversal completes.

Next steps ​

• Page queries, which describe how to fetch one known page.
• Recipes for a complete list-sync workflow.