Launching today
SmartOCR
Multi-provider AI OCR and structured JSON extraction API
1 follower
Multi-provider AI OCR and structured JSON extraction API
1 follower
SmartOCR is a unified AI document extraction API for developers. Pass any PDF, invoice, receipt, or scanned form alongside your target JSON schema, and SmartOCR automatically routes it across AWS Textract, Google Cloud Vision, and frontier LLMs with built-in failover—returning clean, strictly typed JSON in seconds. No bounding boxes, templates, or multi-vendor billing required.
Free Options
Launch Team
Dropbox SignSimple, secure eSignatures for the way your team works.
Promoted
Hey Product Hunt! 👋
Following the launch of SmartOCR Read (our no-code PDF-to-Excel web app), I'm excited to share the developer platform that powers it all: SmartOCR (https://smartocr.dev) — a unified, multi-provider AI OCR and structured JSON extraction API built for developers.
Why we built it:
When building document pipelines for invoices, bank statements, bills of lading, or intake forms, developers usually get stuck choosing between raw OCR engines (which break on complex layouts) or raw vision LLMs (which hallucinate numbers or fail during provider outages).
What SmartOCR does with a single POST request:
Intelligent Auto-Routing & Failover: Automatically routes native PDFs, skewed camera photos, and complex tables across AWS Textract, Google Cloud Vision, and frontier LLMs with zero-downtime automatic failover.
Strict JSON Schema Contract: Pass your exact JSON schema in the request—SmartOCR guarantees the response matches your types and keys without custom regex post-processing.
Zero-Retention Privacy: Documents are processed strictly in-memory and never stored or used to train models.
Simple Pay-As-You-Go Credit Packs: One API key and one unified bill (AWS/Google OCR + LLM costs included) with credit packs from $10 / 100 credits (10¢/page), $40 / 500 credits (8¢/page), and $100 / 2,000 credits (5¢/page)—plus every new developer account gets 10 free API credits on signup (no credit card required) to extract data from your first 10 documents.
Check out the interactive docs at https://smartocr.dev/api-docs — I'd love your feedback and questions on how you're handling document extraction in your stack!