Fieldry

Pre-process

Messy in. Sorted, typed, queued.

Split 200-page scans, classify each document to its schema, and run bulk files at half price. Mindee-style pipeline without the platform lock.

Get Started

Sort first, extract clean

Pick a stage to see what it hands downstream.

pp.1–12 → doc 1invoice

pp.13–40 → doc 2statement

Split

Boundary detection cuts 200-page uploads into single-doc jobs with cited split points.

  • Up to 200pp / 100MB
  • Cited boundaries
  • Mixed scans welcome
Feature 1 of 3

Pre-process

Mixed pile. Sorted queue.

What changes when split, classify and batch run before extraction.

After emphasis: 50%

BEFORE / Manual sort
One schema for mixed files

Statements parsed as invoices

Manual page sorting

Analysts split 60-page PDFs

Sync timeouts on bulk

Large jobs die at 25s

Peak pricing on backlogs

No cheap lane for volume

AFTER / Fieldry pipeline

Split + routed

Fieldry

Auto

Typed per page

Fieldry

Cited

Async with webhook

Fieldry

5GB

Batch at −50%

Fieldry

Delayed
Split and classify included. Batch saves 50%.

Inside Pre-process

SplitSplit multi-doc PDFs before extraction.Explore ClassifyRoute doc types to the right schema.Explore BatchAsync queue for bulk jobs at 50% discount.Explore

Pre-process questions, answered

The questions teams ask most about pre-process.

Talk to us

Whenever one upload holds many documents: a 60-page loan file, a mixed AP inbox, or nightly bulk. Split finds boundaries, Classify routes each piece to its schema.

200 pages and 100MB per file for split/classify, 5GB for batch pipelines. Async jobs callback by webhook.

Delayed queue: jobs run when capacity is free. Same output, same evidence — half the credits. Ideal for backlogs and nightly runs.

Sort every pile first

Split, classify, batch by default.

A 60-page loan file lands mixed. Split cuts it, classify types it, extraction just works.

Request a DemoSee pricing