Document processing pipeline architecture

Files land in storage, an agent extracts structured fields, a person approves anything low-confidence, and the result is written to the database.

AWSocrextractioninvoicesapproval
Upload bucket100 GB standardNew objectS3 event notificationExtraction queue5M requestsExtract fieldsStructured extractionagainst a schemaResolve mismatchesOnly when confidence islowApproval gateguardrailerp_writeWrites the record to theERPBudget capguardrailArcFoundry runtimeruntimeRecords databasedb.t4g.large ·single-AZ · 100 GBProcessed archive1 TB deep archive
11 nodes — 4 infrastructure and 7 agent, on one graph. Drawn by the same layout the product uses.
What is in it

Every resource, and what it costs.

Projections from August 2026 list prices for always-on resources. Connect an account and these become the figures your provider actually bills.

Resources in the Document processing pipeline template with projected monthly cost.
NodeTypeWhat it isProjected
Upload bucketStorage100 GB standard$3.00/mo
Extraction queueMessaging5M requests$2.00/mo
Records databaseDatabasedb.t4g.large · single-AZ · 100 GB$105/mo
Processed archiveStorage1 TB deep archivefrom $4.00/mo
New objectTriggerS3 event notification
Extract fieldsAgent stepStructured extraction against a schemabilled per token
Resolve mismatchesAgent stepOnly when confidence is lowbilled per token
Approval gateGuardrailLow-confidence extractions need reviewenforced at runtime
erp_writeToolWrites the record to the ERP
Budget capGuardrailHalts runs over $250/moenforced at runtime
ArcFoundry runtimeRuntime

Agent steps are priced from provider-reported token usage once the agent runs, not estimated. Guardrails cost nothing and are the reason a runaway agent cannot. One row shows from because those services bill by usage, so the total is a scenario at the stated volumes, not a quote.

Open Document processing pipeline on the canvas.

It loads as an editable graph. Connect an account or instrument an agent and the projected figures above become measured ones.