Search → details (chained)

The hard way: search, then fan out one detail call per zpid. The easy way: POST /v1/search/with-details. We do the chaining server-side.

const api = "https://api.zillapi.com";
const headers = {
  "authorization": `Bearer ${process.env.ZILLOW_API_KEY}`,
  "content-type": "application/json",
};

// 1. Kick off the chained job.
const start = await fetch(`${api}/v1/search/with-details`, {
  method: "POST", headers,
  body: JSON.stringify({
    filters: { status: "for_sale", location: "Austin, TX", beds: { min: 3 } },
    extractionMethod: "PAGINATION",
    maxItems: 200,
  }),
}).then(r => r.json());

const jobId = start.data.job_id;

// 2a. Recommended: register a webhook (see /webhooks-guide/).
//     Skip the polling below.
//
// 2b. Or poll.
async function awaitJob(id) {
  for (;;) {
    const r = await fetch(`${api}/v1/jobs/${id}`, { headers }).then(r => r.json());
    if (["succeeded","failed","timed_out","aborted"].includes(r.data.status)) return r.data;
    await new Promise(r => setTimeout(r, 5000));
  }
}
const finished = await awaitJob(jobId);
if (finished.status !== "succeeded") throw new Error(finished.error);

// 3. Fetch results — these are detail rows, not search rows.
async function* pages() {
  let offset = 0;
  for (;;) {
    const r = await fetch(
      `${api}/v1/jobs/${jobId}/results?limit=500&offset=${offset}`, { headers }
    ).then(r => r.json());
    yield r.data;
    if (!r.meta.has_more) break;
    offset += r.data.length;
  }
}
for await (const batch of pages()) {
  for (const home of batch) {
    console.log(home.zpid, home.address, home.price, home.priceHistory?.length);
  }
}

Cost

Chained jobs charge once per search row + once per detail row. A 200-result search → 200 detail enrichments = 400 units. Save units by running detail extraction only on the zpids you actually need.