TaxHacker
A self-hosted AI accounting app that extracts data from receipts, invoices and PDFs with LLMs and stores it in a structured, categorized database.
- GitHub stars
- 6.7k
- Last commit
- 4 days ago
- Latest release
- v0.8.5
- Licence
- MIT
- Self-hosted
- Yes
TaxHacker is a self-hosted accounting app aimed at freelancers, indie hackers and small businesses that want to spend less time tracking expenses and income. You upload photos of receipts, invoices or PDFs, and an LLM extracts the details, including product names, amounts, items, dates, merchants and taxes, into a structured, spreadsheet-like database.
Custom fields can be defined with your own AI prompts to extract any extra information, and transactions are automatically sorted into categories. The app converts currencies, including crypto, using historical exchange rates from the transaction date, and offers filtering, multi-project support, import and export and custom categories to simplify reporting and tax filing. Topics mention OCR, Gemini, OpenAI and Ollama as model options.
Unsorted documents can be stored until you are ready to process them manually or with AI. TaxHacker is written in TypeScript and licensed under MIT, and the README warns that it is in early development and should be used at your own risk. It suits people comfortable self-hosting who want AI help with bookkeeping.
Key features
- AI extraction from receipts and invoices
- Custom fields defined by AI prompts
- Automatic transaction categorization
- Historical currency conversion including crypto
- Support for multiple projects
- Import and export of data
- Works with OpenAI, Gemini or Ollama models
Pricing: Free and open source under the MIT licence; AI model providers may charge separately.
