Getting started

Quickstart

Create your first extraction. This queues a Blinkit product search for a pincode, then fetches the structured results once it finishes.

1 · Create the extraction

Extractions run in the background, so a create returns 202 with a job id and status: "queued".

curl https://api.auracrawl.com/v1/extractions \
  -H "X-Api-Key: $AURACRAWL_API_KEY" \
  -d '{
    "connector": "blinkit",
    "operation": "search.products",
    "input": { "query": "cold brew", "pincode": "560001" },
    "collect": { "mode": "limit", "limit": 50 }
  }'

import os
from auracrawl import Aura

aura = Aura(api_key=os.environ["AURACRAWL_API_KEY"])

job = aura.extractions.create(
    connector="blinkit",
    operation="search.products",
    input={"query": "cold brew", "pincode": "560001"},
    collect={"mode": "limit", "limit": 50},
)
records = job.wait().records   # polls until the job finishes
print(records)

202 Accepted

{
  "id": "ext_9f2a1c",
  "status": "queued",
  "connector": "blinkit",
  "operation": "search.products",
  "request_id": "req_a1b2c3"
}

2 · Poll until it finishes

Check the job until status is succeeded. See job status for every state.

cURL

curl https://api.auracrawl.com/v1/extractions/ext_9f2a1c \
  -H "X-Api-Key: $AURACRAWL_API_KEY"

200 OK

{
  "id": "ext_9f2a1c",
  "status": "succeeded",
  "counts": { "collected": 50 },
  "usage": { "credits": 150, "tier": "rendered" },
  "request_id": "req_a1b2c3"
}

3 · Fetch the records

Once it has succeeded, page the results by cursor. Each record is one product in your schema.

cURL

curl https://api.auracrawl.com/v1/extractions/ext_9f2a1c/records \
  -H "X-Api-Key: $AURACRAWL_API_KEY"

200 OK

{
  "records": [
    {
      "name": "Blue Tokai Cold Brew Can 200ml",
      "price": 120,
      "mrp": 150,
      "in_stock": true,
      "eta_minutes": 11
    }
  ],
  "next_cursor": null
}

Shortcut. For a small test you can skip the polling: add "delivery": { "type": "sync" } to the create call and the records come back inline (jobs up to 1,000 records). For anything larger, or for unattended pipelines, use the queued flow above or a webhook.