Guides
Each design decision behind the API gets its own page here, not just a line in the Reference — async task model, webhooks, idempotency, pagination, error handling, rate limits, bbox (paid opt-in), and how account-level settings interact with API calls.
Async Task Model
How document and batch processing works asynchronously in the v1 API — the queued/processing/succeeded/failed/canceled state machine and the created_at/started_at/completed_at timestamp chain.
Webhooks
Register a callback URL to get notified when a batch finishes instead of polling — Standard Webhooks signature verification, the event payload shape, retry behavior, the bbox completion event, and how reprocessing a batch gets notified again automatically.
Idempotency
Use the Idempotency-Key header to safely retry document uploads, batch processing, and bbox triggers without paying for or executing the same operation twice.
Pagination
List endpoints in the v1 API use opaque cursor-based pagination — not page numbers — via a next_page_token you pass back verbatim, never decode or construct yourself.
Error Handling
Every v1 error is a JSON object with type, code, message, and doc_url — this page documents the full error type taxonomy and every individual error code, and is exactly what a returned doc_url links back to.
Rate Limits
v1 rate limits apply per account, not per IP address, and every response carries X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers so you can self-throttle before you hit a 429.
Bounding Boxes (bbox)
bbox is an optional, separately-billed second pass that locates exactly where on the page each extracted value came from — model-direct coordinates with no pixel-level verification, so treat precision as approximate, not exact.
Account Settings and API Behavior
v1 reads your account's web-app settings (bbox auto-annotate, retention policy, quality) directly rather than exposing a separate API config layer — this page explains exactly which settings the API always follows and the one it lets you override per request.