Get started

Why EasyOCR exists

EasyOCR fixes a common OCR pain point: most tools flatten everything into plain text. We pair DeepSeek multimodal models with a managed SaaS workflow—batch uploads, Markdown exports, real-time progress—so teams rebuild structured documents without maintaining infrastructure.

Key features

  • Markdown & plain text exports that keep headings, lists, and tables intact
  • Batch uploads up to 50 files with real-time progress updates
  • Results remain in-session only with automatic cleanup for privacy-first workflows

Technology stack

  • Frontend: Astro, React islands, Tailwind, Zustand state
  • Backend: FastAPI, asyncio workers, Pydantic validation
  • OCR Engine: deepseek-ocr.rs wrapping DeepSeek-VL2

Team enablement

  • Enterprise plan unlocks higher quotas, priority queues, and Webhook callbacks
  • Custom prompt templates, export formats, and audit logging available on request
  • Chat with us to scope deeper integrations or bespoke compliance requirements

Quick start guide

  1. Upload PDFs, PNG, JPG, or WEBP files from the home page—batch uploads welcome.
  2. Select an output format and hit “Start recognition” to trigger the background worker.
  3. Preview, copy, or download Markdown results once each file is processed.
  4. Need automation? Enterprise customers can trigger Webhooks or API exports to downstream systems.

Roadmap

Upcoming work includes richer prompt templates, an inline Markdown editor, and optional exports to HTML and Notion. Share your ideas via GitHub Issues—we iterate with the community.