TaxHacker eliminates manual expense tracking for freelancers and small businesses by using large language models to automatically process financial documents. Upload a receipt photo, invoice PDF, or bank statement, and the AI extracts amounts, dates, merchant names, tax information, and product details into a structured database. The tool handles documents in any language, processes handwritten receipts, and works with multi-currency transactions across 170+ world currencies and 14 popular cryptocurrencies including BTC, ETH, and DOT.
What sets TaxHacker apart is its flexibility in both AI and data modeling. You can connect your preferred LLM provider—OpenAI GPT, Google Gemini, Mistral, or run completely offline with Ollama and LM Studio. Beyond standard extraction, you can define unlimited custom fields with tailored AI prompts: a law firm might extract client case numbers from invoices, while an e-commerce seller could pull product SKUs to calculate per-item costs. Multi-project accounting lets you track expenses across different clients or business lines with separate reporting and CSV export for tax filing.
Built with Next.js and PostgreSQL, TaxHacker deploys via Docker Compose on any infrastructure you control. Created by vas3k, a well-known developer and blogger, the project is actively maintained under the MIT license with a growing community. All financial data stays on your server—no cloud processing, no vendor lock-in, and no subscription fees beyond whatever LLM API costs you choose to incur.