Document-aware
Supports a generic, versioned document-type catalog without application-specific tables.
01 Document intelligence infrastructure
Production environment
SCE preserves each source, records its ownership, and returns a durable capture receipt. Document intelligence remains outside version 0.1.
01AcceptOne source per request
02PreserveImmutable original in MinIO
03QueueStops at awaiting_engine
02 Live capture intake
Real files are sent as multipart form data to POST /api/v1/capture. The demonstration button never sends a request.
Selecting one file sends a multipart POST to /api/v1/capture.
03 Canonical pipeline
Version 0.1 completes upload and original storage, queues image processing, and stops before any OCR or AI runs.
OCR, classification, AI, rules, confirmation, output, and integration remain unimplemented framework stages.
04 More than OCR
Supports a generic, versioned document-type catalog without application-specific tables.
Uses stable REST contracts and preserves the ownership context needed by downstream systems.
Computes SHA-256, invokes the scan hook, encrypts metadata, and writes an append-only audit event.
Stores immutable file objects outside PostgreSQL with versioned MinIO or S3-compatible identity.
05 Independent plugins
Applications integrate through the same tenant-scoped API while product knowledge remains outside the SCE core.
curl -X POST https://capture.combridgecorp.com/api/v1/capture \
-H "Authorization: Bearer $SCE_API_KEY" \
-H "Idempotency-Key: sce-web-unique-request" \
-F "file=@invoice.pdf" \
-F "schema_id=document.generic.v1" \
-F "document_type_key=unknown"