Where this sits in the flow
What it Does

Use whichever fits the job: a quick poll for small files, a webhook for volume, or a CSV and JSON handoff when a person is in the loop.
Async by default

Every extraction runs as a background job. Submit, receive a job ID, and either poll for status or wait for the webhook. It is the same path whether the document is one page or two hundred, so your integration never needs a special case for big files.
Which Engine
Queued
Processing
Complete
Delivered
job_9f2a queued, 62 pages
10:41:02
Processing with Contextual AI
10:41:03
Complete, 14 fields extracted
10:43:18
Webhook delivered to yourapp.com
10:43:18
Field definitions as code

Build a field set in the interface, using the auto-generated prompts, then export it as JSON. That same file drives the API, so the person who understands the documents and the person who writes the integration are working from one definition.
Where the data goes

Delivery
Result format
JSON
Webhook
yourapp.com/hooks/done
Also available
CSV download
Tracking
Documents view
Keep Exploring

Standard Layout
Clean, digital documents
Explore Products
Advanced Vision
Scans, photos, complex layouts
Explore Products
Form Intelligence
Fields, checkboxes, tables
Explore Products
Contextual AI
Meaning-aware extraction
Explore Products
See the full Extractor Studio overview