Multi-engine OCR
Self-hosted PaddleOCR for cost efficiency, or layer in Azure Document Intelligence, Google Document AI, and vision-LLMs (Qwen2.5-VL, Vintern, DeepSeek) for specialized document types.
OCR · Document Parsing
Turn any document into structured data.
A self-hosted, multi-engine OCR platform that parses PDFs, scans, and office files into machine-readable data. Layout analysis, table recognition, and word-grounded LLM extraction into your own schemas — with GPU acceleration and no per-page vendor fees.
Multi-engine · 80+ languages · GPU-accelerated
PaddleOCR
Azure Document Intelligence
Google Document AI
Vision-LLMs
GPU-accelerated
Capabilities
A full pipeline from pixels to structured data
Multi-engine OCR, layout and table understanding, and schema-driven extraction — all behind one unified REST API.
Layout & table recognition
PaddleLayout v3 detects headers, footers, and regions; SLANet and TSR engines extract tables with cell boundaries and intelligent multi-page merging into HTML or structured cells.
Schema-driven extraction
Define JSON Schemas for your fields and extract typed data — strings, numbers, dates, objects, arrays — with LLM extraction that understands context and relationships.
Word-grounded results
Every extracted field carries a confidence score and precise word-level bounding boxes — perfect for trust-but-verify review UIs and high-fidelity data workflows.
Sync & async APIs
Immediate /parse and /extract for small docs, or enqueue large jobs with /parse_async and poll status — no client-side timeouts on bulk processing.
GPU-accelerated
GPU acceleration for fast, cost-effective processing at scale — with automatic CPU fallback so it runs anywhere.
The pipeline
Modular stages, one clean result
A per-page middleware pipeline — raster → layout detection → OCR → table recognition → block assembly — that you can inspect, customize, and test stage by stage.
- Unified structured format inspired by Azure Document Intelligence — words, paragraphs, tables, and figures tied to one content string via UTF-16 span offsets.
- Multiple output formats — JSON, Markdown, or plain text via content negotiation.
- Chunking for RAG — token-bounded or semantic (paragraph / table / image) chunking for embeddings and AI pipelines.
- 80+ languages via PaddleOCR, with native Vietnamese optimization out of the box.
Raster · 300 dpi
page 1 of 12
Layout · 6 regions
PaddleLayout v3
Assembled · 1 table, 4 ¶
UTF-16 spans mapped
Own your OCR
Self-hosted. No per-page tax.
Self-hosted PaddleOCR eliminates vendor lock-in and per-page OCR costs while holding 95%+ accuracy on English and Vietnamese — and you can still reach for cloud engines per request when a document demands it.
- Pluggable backends — the same API dispatches to PaddleOCR, Azure, Google, or vision-LLMs; switch per document type, not per contract.
- Native multi-tenancy — documents, jobs, and API keys are org-scoped from day one, ready for SaaS and white-label.
- Runs anywhere — deploy on your own servers or cloud, with an optional GPU mode for high-volume workloads.
- Included React dashboard — uploads, parse/extract previews, API-key management, and job monitoring.
PaddleOCR · self-hosted
default · $0 / page
Azure / Google · optional
per-request routing
Vision-LLM · edge cases
Qwen2.5-VL · Vintern
Why Reflexify OCR
Accuracy you can verify, costs you control
Trust-but-verify UX
Word-grounded fields with confidence scores and bounding boxes power precise correction workflows — humans review only what needs it.
No layout/table vendor deps
Layout detection and table recognition are built into one pipeline — from PDF to structured data without stitching separate services.
Bulk-safe by design
A built-in async job queue and worker pool remove client-side timeout concerns for high-volume document processing.
Multi-tenant, not bolted on
Org-scoped isolation is native to the schema — enabling SaaS deployments and embedded white-label scenarios out of the box.
Use cases
Wherever documents pile up
Invoice & receipt processing
Extract line items, amounts, dates, and vendor details from hundreds of invoices a day via async jobs, with word-grounded review for exceptions.
Mortgage & loan documents
Auto-parse applications, income verification, appraisals, and titles. Schema-driven extraction feeds structured data straight into underwriting.
Legal document discovery
Bulk-parse contracts, transcripts, and filings; use table recognition and hierarchical chunking for semantic search and RAG.
Healthcare records
Convert scanned charts, lab results, and prescriptions into structured EHR data with multi-language support and high accuracy.
RAG & knowledge bases
Feed clean, chunked document content into embeddings pipelines for AI assistants and semantic search across your corpus.
Signing prep & automation
Pair with AgentFlow and eSign — auto-extract signer fields and form regions to stage documents for signature without manual prep.
Stop retyping your documents.
Self-hosted, multi-engine OCR and extraction with word-grounded accuracy and no per-page fees. See it run on your documents.