Upload Center
Real intake: OCR scanning for documents, plain storage for attachments. Nothing is posted to the records from here.
Scan a document (OCR)
PNG, JPEG, BMP or PDF up to 30 MB. A PDF's own text layer is read directly; scanned pages and images go through Tesseract OCR. Fields are extracted with the engine's confidence — a field it could not find has no score at all.
Drop a document to scan
PNG, JPEG, BMP or PDF · ≤ 30 MB
Document type: invoice — Only the “invoice” field schema is configured for this company — other uploads still capture text but extract no fields.
Job history & review
Every scan is listed newest-first. A job can only be drafted once no field still needs review — the draft is a proposal, never a posted entry.
Loading jobs…
Pick a job to see its fields and corrections.
Attachments — store a file
Video (mp4, mov, m4v, webm, mkv), images (png, jpeg, webp) and PDF up to 100 MB. Stored means stored — no analysis runs on these files; a video here plays back, nothing more.
Drop a file to store
video mp4/mov/m4v/webm/mkv · png/jpeg/webp · pdf · ≤ 100 MB
Loading attachments…
Production remains dependent on a configured external object store. A scan never creates or posts an invoice or journal entry automatically. The scan reports document-level engine confidence; any field scores are engine-reported, not independent per-field confidence measurements.