⌕ soon
Account Online
Junbi Engine · document preparation core

Capture

Open a source document locally, preview it on-device, then explicitly send it to Junbi for request-scoped extraction and normalization. NexoTori does not retain the source file or the processing result on the server.

Processing limits: 20 MB, 100 PDF pages, 30 million image pixels. Split or downsize larger documents. Upload buffers stay in server memory.

No user file repository: preview stays in this browser. Processing occurs only after you click “Process with Junbi,” and the server handles the file for that request only.
Source

Document preview

No file
⇧
Drop a document here or use “Choose file” above PDF, DOCX, images, text, CSV/TSV, JSON · maximum processing size 20 MB
Local workspace

Document details

No persistence
Name
—
Type
—
Size
—
Last modified
—
Server upload
Not requested
Server persistence
Disabled by design
Junbi v0.2

Native preparation providers

RapidOCR is now the offline/default OCR provider. PDFs use a hybrid strategy: embedded text when reliable, OCR when a page looks scanned, or forced OCR when requested.

Analytics preferences

Optional Google Analytics helps us count visits to this site. It uses browser identifiers and device information. We send a fixed site name and homepage address, not your account ID, page URL, document contents, search terms or form entries. Advertising features are disabled. Your choice applies to this browser on this site.

Analytics is off.