Local-First Payroll & Resident Tax Automation

The description states that this is a local-first payroll application for Japanese small businesses, designed to automate parts of the resident tax notice processing workflow. The author reports building a Windows desktop app using Python and local storage technologies (SQLite, Streamlit) that reads tax notices from PDFs or scanned images, extracts data via OCR, separates multi-employee records, and links confirmed tax amounts to payroll calculations without cloud transmission.

The project is self-reported as a hackathon submission with no evidence of revenue, customers, or traction. The author describes building the tool for personal use and small business operations but does not provide any information about actual deployment, usage, or adoption by others.

The description states that the product is a local-first Windows desktop application for processing Japanese resident tax notices. It imports PDFs or scanned documents (JPG/JPEG/PNG), uses OCR as fallback, separates multi-employee records, extracts annual and monthly tax amounts, allows human review, stores revision history, links confirmed amounts to payroll calculations, and generates payslips and accounting CSV files.

The application is built using Python with Streamlit, SQLite, ReportLab, Tesseract OCR, and other libraries. It is designed to process sensitive payroll data locally without sending it to the cloud.

Evidence The author's own write-up describes the functionality and technical stack used.

The description states that this tool addresses a manual, error-prone process where Japanese small businesses must manually read resident tax notices, break down annual amounts by month, match them to employees, and enter into payroll. The author positions it as connecting fragmented back-office tasks while keeping sensitive data local.

The product is described as a practical tool for small-business payroll staff that reduces time spent on manual entry and human error.

Evidence The author's own write-up describes the problem being solved and the intended value proposition.

The description states that this tool targets Japanese small businesses that receive resident tax notices. These are described as businesses where payroll staff must manually process tax notices, which is time-consuming and error-prone.

Evidence The author's own write-up describes the target use case and customer pain points.

Not evidenced.

Evidence No information provided about pricing, monetization, or business model in the description.

The description states that the application is built as a local Windows business application using Python, Streamlit, SQLite, ReportLab, Tesseract OCR, pypdfium2, pypdf, and Pillow. It uses Codex for development assistance and implements business rules based on real small-business payroll operations.

Evidence The author's own write-up describes the technical stack and development approach.

Not evidenced.

Evidence No information provided about actual deployment, usage, customers, or traction beyond the author's personal account of building it for a hackathon.

Not evidenced.

Evidence No mention of competitors or market context in the description.

The description states that this is a single-person project built as a hackathon submission. There is no evidence of:

• Actual deployment or usage by small businesses
• Integration with existing payroll systems
• Support for multiple municipal notice layouts
• Validation of tax calculations
• Security testing or compliance considerations
• Scalability beyond one user
• Long-term maintenance plan

The author notes challenges around OCR accuracy, employee matching, and database migration, suggesting potential technical limitations.

Evidence The author's own write-up describes the project as a personal hackathon effort with no evidence of broader adoption or commercial viability.

1. Has this tool been tested with actual payroll staff from Japanese small businesses?
2. What is the accuracy rate of OCR and employee separation in real-world documents?
3. How does the system handle variations in municipal notice layouts beyond those already supported?
4. Are there any known compliance or regulatory issues with how this handles tax data?
5. What are the specific business rules that govern how tax amounts are applied to payroll calculations?
6. Has the author considered integration with existing payroll software used by small businesses?
7. How does the system handle edge cases like missing or corrupted documents?
8. What is the expected time savings for users compared to manual processing?