DO Patent logo

DO Patent

Extract, verify, and edit molecular structures from PDFs with >98% accuracy — 12x faster than traditional OCR and sketch-to-SMILES tools.

Solution by Deep Origin
Visit website

Overview

DO Patent, developed by Deep Origin, is a molecular structure extraction platform designed for researchers, academic labs, independent scientists, and life sciences organizations that need to rapidly convert PDF documents — including patents and scientific publications — into ready-to-use molecular data. With greater than 98% full-molecule image data extraction accuracy and extraction speeds 12 times faster than competing tools, DO Patent eliminates the time-consuming process of manually redrawing chemical structures.

By uploading any PDF, users can instantly convert embedded full-molecule structures into verified SMILES strings. DO Patent's extraction engine is built to outperform traditional OCR tools and generative AI sketch-to-SMILES approaches, enabling scientists to focus on analysis rather than data preparation. The platform also allows users to build unique SMILES datasets from sources that major databases cannot legally or technically aggregate, giving organizations full control over their competitive intelligence.

Core Extraction and Verification Capabilities

  • Automatic full chemical structure extraction from PDFs, including patents and scientific publications
  • Greater than 98% full-molecule image data extraction accuracy
  • Extraction speeds 12 times faster than competing solutions
  • Each extracted molecule is assigned a confidence score to help users quickly identify and correct uncertain predictions
  • Medium-confidence structures are flagged with direct links to the exact source image and page for rapid verification and correction
  • Full data provenance and transparency to support reproducible science

Editing and Data Management Features

  • Integrated molecule visualizer and editor for refining structures directly within the interface
  • Bulk PDF uploads for high-throughput processing
  • Bulk SMILES export, including Balto-ready formatting
  • Ability to build custom SMILES datasets from sources unavailable in major databases
  • Private, secure processing and storage of all molecular data

Available Plans

  • DO Patent Individual: Designed for academic labs, independent researchers, and early-stage biotechs, this plan provides automatic full chemical structure extraction, an integrated molecule visualizer and editor, full data provenance and confidence scoring, bulk PDF uploads, bulk SMILES export, and private secure processing. A free tier is available.
  • DO Patent Enterprise: Built for organizations that require DO Patent capabilities at scale, with tailored options available by contacting Deep Origin directly.

DO Patent is developed by Deep Origin, headquartered in South San Francisco, CA. The platform is designed with security and data privacy in mind, offering private and secure processing and storage for all user data.

Meta

Domain
Research Intelligence & Discovery
Subdomain
Chemical & Patent Intelligence
Software type(s)
Analytical Platform
Deployment type(s)
Cloud / SaaS
Industry vertical(s)
Academic / ResearchBiotechCROPharma
Development stage(s)
Research & DiscoveryPreclinical / Pre-Market
Target user(s)
Research ScientistBioinformatician / Computational ScientistMedicinal Chemist