idpura — SaaS Platform for Data Extraction from Documents

The Challenge

A data analyst at an accounting firm in Valencia. Monday morning. Forty-seven supplier PDFs piled up on the desk. Each one filled with tables, amounts, dates, and references that need to be manually copied into a spreadsheet.

Three hours later, the spreadsheet has errors in row 38 because a decimal ended up where it shouldn’t. Start over.

That’s the daily reality for accounting firms, law offices, and Business Intelligence teams across Spain. Copying and pasting data from documents is slow, error-prone, and doesn’t scale. The existing solutions either rely on expensive AI with unpredictable results or require technical setups that an accounting team isn’t going to build.

We needed to build something that worked from minute one. No configuration. No errors.

The Solution

idpura is an in-house product by It’s Genki. We designed it, we developed it, and we operate it. It’s not a client project — it’s the proof of what our tech stack can do when pushed to the limit.

Full web application with Astro 5 and React 19 — the frontend combines Astro’s speed for marketing pages (home, pricing, documentation) with interactive React components for the application where users upload, process, and download documents. Two worlds under a single domain.

Async batch processing — documents enter a queue and are processed in the background. The user sees progress in real time without the browser locking up. Up to 500 files per job. Average time per document: under 2 seconds.

Professional authentication — Google sign-in and organization support for teams. Every application route is protected. This isn’t a toy login.

GDPR-compliant by design — processing on dedicated European servers. Files are automatically deleted within 24 hours. No dependency on third-party clouds.

Credit-based pricing model — 4 plans starting at 9 euros per month. Each extraction module consumes credits per page. Credits never expire. Users know exactly how much they’ll pay before processing a single document.

Bilingual from day one — Spanish and English with the same i18n architecture we use across all our Astro projects.

The Result

idpura is in production and in open beta. The numbers so far:

• 2.4 million pages processed
• 98.7% accuracy in structured text extraction from native PDFs
• Under 2 seconds per page processing time
• 47 documents in 94 seconds — real data from a demo with accounting firm files

The roadmap includes an AI-powered financial extractor for Spanish invoices, a REST API with OpenAPI documentation, and an organization system for teams.

Why This Matters for Your Project

If you’re reading this and thinking “I need something like that for my business,” that’s exactly what we want you to understand.

It’s Genki doesn’t just build fast websites for local businesses. We design and develop full web applications: dashboards, platforms with login systems, data processing pipelines, integrations with external APIs, and automations with n8n.

idpura is the proof that we have the technical muscle to take an idea from the first wireframe all the way to production with real users.

Technologies Used

Astro 5 · React 19 · PostgreSQL · n8n