Reducto provides an async job-based invocation that lets you submit a massive amount of documents for processing without waiting for each one to finish sequentially. This is ideal for large workloads where you want to queue and be notified of completion later. This applies to all of our endpoints (/parse, /extract, /split).

Using .run_job() has no limit on the amount of documents you can process concurrently.

Each SDK has a .run_job() function that then returns a job_id that can be used with a webhook notification on completion, or poll via client.job.get("<job_id>"). If you’d rather not implement polling or webhooks, see our run() batch processing method.

Make sure you have a Reducto SDK set up, see how in Quickstart.

​
Example

import time
from reducto import Reducto

client = Reducto()

submission = client.parse.run_job(document_url="https://ci.reducto.ai/onepager.pdf")

while (job := client.job.get(submission.job_id)).status == "Pending":
    time.sleep(1)

print(job.result)