Extracta — Self-Hosted AI Invoice & Receipt Data Extractor (BYO Key)
Extract invoice & receipt data with your own AI key — self-hosted, human-reviewed.
The problem it solves
Every business processes invoices and receipts, and someone still retypes the vendor name, date, and line items into a spreadsheet by hand, because the numbers live in a PDF attachment or a phone photo, not in the accounting tool. Black-box SaaS extraction services solve the typing, but they own your documents, charge per-page subscriptions, and require trusting a third party with financial data.
- Manual re-entry of invoice and receipt data
- SaaS tools that store a copy of your financial documents
- Recurring per-page charges for something you could run yourself
What you get
Vision-LLM extraction, your key
Sends each invoice or receipt to OpenAI, Anthropic, or a local Ollama model using your own API key — your documents never leave your server.
Full field extraction
Pulls vendor, invoice number, invoice date, due date, currency, subtotal, tax, total, and full line items.
Arithmetic cross-check
Flags when subtotal plus tax doesn't equal total, or line items don't add up, so you catch model errors before exporting.
Human-in-the-loop review
Review and correct every extracted field on screen before it enters your records.
CSV/JSON export
Export clean, spreadsheet-ready CSV or developer-friendly JSON.
Switch providers anytime
Change between OpenAI, Anthropic, and Ollama in Settings in under a minute.
Pricing — one-time, yours forever
Available soon — get notified
Checkout isn't open yet. Leave your email and we'll send one note the moment Extracta — Self-Hosted AI Invoice & Receipt Data Extractor (BYO Key) can be purchased — nothing else.
Screenshots
Honest limitations
- No direct accounting software integration (exports CSV/JSON, you import manually)
- Single-tenant: one admin account per installation
- No OCR built in — relies entirely on the LLM's vision capability
- Scanned PDFs (images embedded in PDF) may have lower accuracy than native text-layer PDFs
- Extraction accuracy is not guaranteed for any specific format, language, or jurisdiction
2-minute web installer, plain PHP 8+ with MySQL or SQLite, zero Composer dependencies — includes a demo mode with pre-baked sample data.
Frequently asked questions
Is my API key safe?
It's stored in your own database and transmitted only in the outbound HTTPS request you trigger to your chosen provider — never logged, echoed, or passed through any Extracta infrastructure.
Will this run on shared hosting?
Yes — PHP 8+ with pdo, curl, and fileinfo, extensions present on virtually all cPanel/Plesk hosts; installs through the web installer in about two minutes.
Do I need my own API key?
Yes — an OpenAI or Anthropic API key, or a local Ollama instance, is required; Extracta doesn't include one.
What about PDFs?
Image formats (PNG, JPEG, WebP) work reliably with all three providers; Anthropic handles text-layer PDFs natively, while OpenAI and Ollama treat PDFs as best-effort.
What does it cost to use?
Extracta itself is a one-time purchase; your ongoing cost is only your own LLM usage, which varies by provider and model and should be checked with your provider.
See it running before you buy
The live demo is being prepared — check back shortly.
Demo — coming soon