Freelancers and sole proprietors share a universal pain point: tracking business expenses is tedious, error-prone, and always urgent at the worst possible time. The typical freelancer accumulates receipts in a jacket pocket, photographs some of them, loses others, and spends an evening each month manually typing amounts into a spreadsheet. When tax season arrives, the spreadsheet has gaps, the receipt photos are out of order, and the accountant sends back questions about missing entries.
Existing expense tracking apps solve parts of this problem but create new ones. Most require uploading financial data to their servers, which makes freelancers uncomfortable. Many are designed for enterprise teams with features and pricing that solo operators do not need. The ones targeted at individuals tend to be ad-supported or lock export features behind a subscription paywall.
SimplBiz was designed around a specific user: a Slovenian freelancer (s.p.) who photographs receipts on the go, needs mileage tracking for business travel, and sends a monthly expense report to their accountant. Every feature was evaluated against this persona rather than trying to serve every possible business type.
The decision to use Google Drive instead of a traditional database was driven by three factors: privacy, cost, and trust. Storing user financial data in our own database would require GDPR compliance infrastructure, encryption at rest, backup systems, and the ongoing liability of being a custodian of sensitive data. Google Drive eliminates all of this because the data lives in the user's own cloud storage account.
The implementation uses the Google Drive API with the drive.file scope, which grants SimplBiz access only to files it creates rather than the user's entire Drive. On first connection, the app creates a folder called "SimplBiz" with monthly subfolders. Receipt images are uploaded as JPEG files. Expense metadata (amount, date, vendor, category, notes) is stored in a JSON file that SimplBiz reads on launch and writes on every change.
Cost was a significant factor. Google Drive provides 15GB of free storage per account. A typical receipt photo is 200-500KB after compression, and the metadata JSON grows by roughly 200 bytes per expense entry. A freelancer logging 100 expenses per month with receipt photos would use about 600MB per year. At this rate, the free tier would last 25 years. We effectively have zero storage costs, which means we do not need to charge users a subscription to cover infrastructure.
Receipt scanning is SimplBiz's flagship feature. The user photographs a receipt, and Gemini Vision extracts the total amount, date, vendor name, and suggests a category. The prompt is carefully structured to handle the messy reality of receipts: faded thermal paper, handwritten additions, multiple currencies, partially obscured text, and non-standard date formats.
The system prompt instructs Gemini to return a JSON object with specific fields: amount (as a number), currency (ISO code), date (ISO format), vendor (cleaned-up business name), suggestedCategory (from a predefined list), and confidence (0 to 1). The confidence field is critical because it allows the UI to highlight low-confidence extractions that need user verification. A receipt photographed at a bad angle in dim lighting might parse correctly but with 0.4 confidence, prompting the user to review rather than blindly trusting the result.
We tested the scanner against 200 real receipts across four countries and three languages before launch. Accuracy on clear, well-lit photos of printed receipts exceeds 95% for amount and date extraction. Performance degrades on handwritten receipts, receipts with multiple transaction lines (where identifying the total is ambiguous), and receipts in languages with non-Latin scripts. For these cases, the low confidence score triggers manual review, so incorrect data is flagged rather than silently accepted.
Slovenian tax regulations allow sole proprietors to claim mileage deductions at specific rates that differ between daily commute travel and business trip travel. SimplBiz implements both rate categories with automatic round-trip calculation. The user enters a destination and one-way distance, SimplBiz doubles it, applies the correct legal rate, and logs the entry with the calculated deduction amount.
The mileage tracker maintains separate totals for commute and business travel because they have different annual caps and rate structures. These rates change periodically when the government updates them, so they are stored as configuration values rather than hardcoded. When rates change, we push an app update with the new values, and the UI shows which rate was applied to each historical entry so users can verify correctness.
Integration with the expense system means mileage deductions appear alongside receipt-based expenses in the monthly export. The accountant receives a single spreadsheet that includes both types of deductions, each clearly labeled with the applicable rate and regulation reference. This level of detail reduces back-and-forth questions between freelancers and their accountants during tax preparation.
SimplBiz's export feature generates a comprehensive package that accountants can process without additional questions. The export includes a CSV or Excel spreadsheet with every expense for the selected period, organized by date with columns for amount, currency, vendor, category, payment method, and notes. Each row includes a Google Drive link to the corresponding receipt image, so the accountant can verify any entry by clicking the link.
The export pipeline also generates a summary sheet with totals by category, monthly trend comparisons, and budget variance if budgets were set. For mileage entries, a separate section lists each trip with distances, rates, and calculated deductions. The goal is a document that the accountant can file directly rather than reformatting into their own template.
Email integration allows the user to send the export directly to their accountant from within the app. SimplBiz pre-populates the email with a professional summary message and attaches the spreadsheet. For accountants who prefer working directly in Google Drive, the export can alternatively be written to a shared folder that both parties access. This flexibility accommodates different accountant workflows without forcing a specific process.
SimplBiz launched with English and Slovenian language support, which covers the primary target market of Slovenian freelancers who may work in either language. Internationalization was implemented from the start using a key-based translation system where every UI string is referenced by a translation key rather than hardcoded text.
The receipt scanner handles multi-language receipts regardless of the app's UI language. A Slovenian user shopping in Austria receives a German-language receipt, and Gemini Vision correctly extracts the data because the model is multilingual. Currency detection works similarly: the model identifies EUR, USD, GBP, and other currencies from receipt formatting and symbols rather than relying on the app's locale setting.
Adding a new language requires translating approximately 300 string keys and testing the UI for layout issues caused by text length differences. German and Croatian are the next planned languages based on user requests. The translation files are plain JSON objects, making it straightforward for community contributors to submit translations without needing to understand the codebase.
SimplBiz's privacy architecture has three layers. First, receipt images and expense data are stored exclusively on the user's Google Drive. SimplBiz's servers never see or store financial data. Second, the Gemini API calls for receipt scanning use the BYOK model, where the user's own API key is stored in localStorage and sent directly to Google. Third, the app runs as a client-side SPA with no analytics that track financial data, no server-side session storage, and no third-party data processors beyond Google's own APIs.
The practical result is that SimplBiz cannot access a user's financial data even if compelled to. There is no database to query, no server logs containing receipt contents, and no API key storage on our infrastructure. If a user revokes the Google Drive permission, SimplBiz immediately loses all access to their data. This is not just a privacy policy claim; it is an architectural guarantee that can be verified by inspecting the network requests in browser DevTools.
This architecture does create a support challenge. When users report issues with receipt scanning or data corruption, we cannot reproduce the problem with their data. We rely on anonymized error telemetry (error codes, not data contents) and ask users to share example receipts voluntarily. Most are willing to share a problematic receipt photo once they understand that we genuinely cannot see their data without their active cooperation.
SimplBiz found its initial user base through Slovenian freelancer communities on Facebook and Reddit. The privacy-first messaging resonated strongly in these groups, where distrust of expense tracking apps that "store your financial data on their servers" is a common theme. Early adoption was driven almost entirely by word of mouth from users who appreciated that their receipts stayed on their own Drive.
The most impactful feedback was about the onboarding flow. The initial version required users to create a Google AI Studio API key before they could use receipt scanning, which created a significant drop-off. We added a limited free tier that provides 10 AI scans per month through a shared key proxied via a Cloud Function, with the BYOK option for users who need more. This change doubled the completion rate of the onboarding flow.
Feature requests cluster around two themes: recurring expense templates for fixed monthly costs like rent and subscriptions, and integration with Slovenian banking APIs for automatic transaction import. Both are on the roadmap, with recurring expenses scheduled for the next release. Banking API integration is more complex due to PSD2 compliance requirements and the need to work with third-party aggregators, but it represents the natural evolution from manual expense entry to automated financial tracking.