Hello guys, I’ve built an offline AI-powered invoice-processing system and I would really appreciate your feedback.
I built a desktop app called Baraa Invoice System. It processes invoices (PDFs, images, scanned copies, mobile photos) completely offline and turns them into structured data. No cloud, no API calls — everything happens on the user’s machine for privacy.
I’m using a dual-engine setup:
A Vision layer for OCR, layout reading, table detection, skew correction, etc.
A local Logic layer (LLM) to understand the extracted text, separate buyer/seller info, detect line items, validate totals, fix misreads, and flag errors.
The app has two modes:
Single-invoice workbench where users can manually verify with a side-by-side preview.
Batch mode where it processes hundreds of invoices automatically with status indicators.
It also auto-renames files, organizes folders, validates calculations (Qty × Price), checks tax/subtotal consistency, and exports to Excel, CSV, JSON, or SQLite.
My main goals are:
Work fully offline
Be accurate across different invoice formats
Be simple for non-technical users
Provide reliable structured output for businesses
I’m still improving the model, logic rules, and file-handling flow.
If any developers here can give feedback — UX issues, architecture mistakes, performance ideas, or anything I should improve — it would help me a lot.
You can check it out here: https://invoicesystemai.com
Now I am in a situation of selling this software ownership to someone.
I want to open source this software but I have to pay my college fees so I am selling. people asking for 1k - 2k dollars. is it reasonable price?
and I am struggling to sell it can anyone guide me.
thank you guys
Top comments (0)