Routes
/ocr/editor-soft?ocr_session=[id]Flow 8
Photo → OCR → Editor
Take a photo (or a series of photos) of a physical document, upload them, and BrieflyGo will read the text with OCR, correct errors with AI, detect fillable fields, and open the result in the Soft Editor — ready to edit, annotate, sign, and send.
Open screenshot full size
ELI10
Take a photo (or a series of photos) of a physical document, upload them, and BrieflyGo will read the text with OCR, correct errors with AI, detect fillable fields, and open the result in the Soft Editor — ready to edit, annotate, sign, and send.
W5
Who: anyone with a paper document (contract, form, invoice, receipt) who needs a digital editable copy.
What: photo upload, EasyOCR recognition, AI text correction, fillable field detection with bbox coordinates, editor handoff.
When: before digitising, editing, signing, or archiving a physical document.
Where: /ocr upload page → background processing → /editor-soft?ocr_session=[id].
Why: paper documents should enter the same risk/edit/send workflow as digital ones without any manual retyping.
Important buttons
Watch-outs
Supports JPEG, PNG, WEBP, HEIC/HEIF (iOS). Up to 10 photos, 15 MB each.
Processing is async — the frontend polls status every 2 s until done (max 6 min).
OCR fillable fields are injected into the editor only after loadFile() completes via pendingOcrFieldsRef.
The AI correction model (default openai/gpt-oss-120b:free) is configurable in Admin → AI Models → OCR Text Correction.