AI-Based Receipt (Slip) Reading & Financial Tracking Platform (Mobile + Web)
Budget: €3,000 – €5,000 EUR
I need a working proof-of-concept that lets me upload receipts, invoices, and bank statements and instantly turns them into structured records. As soon as a file lands in the system, an OCR engine (Google Vision, AWS Textract, Tesseract, or a stack you recommend) should detect the date, total amount, vendor name, and individual item descriptions, then write everything into a database. CSV or JSON exports can come later—right now the goal is a clean relational schema I can query.
A minimal web interface is fine: a drag-and-drop zone or file picker, a thumbnail/preview of the document, and a simple table showing the parsed results. Code quality, clear separation between the OCR module and the persistence layer, and straightforward deployment (Docker or similar) matter more than pixel-perfect UI.
Acceptance criteria:
• Upload image or PDF versions of receipts, invoices, and bank statements.
• Automatic extraction of date & amount, vendor information, and item descriptions with solid accuracy on legible documents.
• Parsed data saved to a database and immediately visible in the UI.
• Setup instructions and containerised build so I can run the MVP locally or on a small cloud instance.
If you have prior experience with document OCR, data extraction, or fintech tooling, let’s talk architecture and timeline.
A minimal web interface is fine: a drag-and-drop zone or file picker, a thumbnail/preview of the document, and a simple table showing the parsed results. Code quality, clear separation between the OCR module and the persistence layer, and straightforward deployment (Docker or similar) matter more than pixel-perfect UI.
Acceptance criteria:
• Upload image or PDF versions of receipts, invoices, and bank statements.
• Automatic extraction of date & amount, vendor information, and item descriptions with solid accuracy on legible documents.
• Parsed data saved to a database and immediately visible in the UI.
• Setup instructions and containerised build so I can run the MVP locally or on a small cloud instance.
If you have prior experience with document OCR, data extraction, or fintech tooling, let’s talk architecture and timeline.