Turn Any PDF into Ready-to-Use Molecules In Minutes
Extract, verify, edit, and export molecular structures from patents and publications with >98% accuracy and 12x faster extraction than competitors.
Why DO Patent?
Spend Time Analyzing, Not Redrawing
Upload PDFs to instantly convert embedded molecular structures into verified SMILES strings. Outperforms traditional OCR and GenAI sketch-to-SMILES tools.
Trust Every Atom
Each extracted molecule carries a confidence score. Medium-confidence structures are flagged with links to source images, enabling quick verification and correction.
Your Data, Your Way
Refine, edit, and export molecular data within the interface. Build unique SMILES datasets from competitive sources while maintaining full control over your data.
Tested on Real Pharmaceutical Patents
Over 99% of full-molecule structural elements correctly extracted across 30+ patents and thousands of structures. Manually validated bond-by-bond by an experienced chemist.
| Patent ID | Drug | Company | Molecules | Overall Accuracy | High-Confidence |
|---|---|---|---|---|---|
| US7838499 B2 | Brenzavvy | Theracos | 292 | 99.1% | 97.4% |
| US2022/0324863 A1 | LXE408 | Novartis | 526 | 98.4% | 99.8% |
| US9447106 B2 | Brukinsa | BeiGene | 732 | 99.0% | 99.7% |
| US8410103 B2 | Cabenuva | Shionogi | 260 | 98.3% | 86.6% |
Single atom or bond errors marked as failed extractions. Task requiring 100+ hours of manual validation completed in minutes with DO Patent.
Read more about our validation and why we built DO PatentSimple, Transparent Pricing
DO Patent Individual
- Automatic full chemical structure extraction
- Integrated molecule visualizer and editor
- Full data provenance and confidence scores
- Bulk PDF uploads
- Bulk SMILES export
- Private, secure processing
DO Patent Enterprise
Custom solutions for teams and organizations with high-volume extraction needs, API access, and dedicated support.
Contact UsFrequently Asked Questions
What file formats does DO Patent support?
DO Patent currently supports PDF files, including scanned documents and native PDFs from patents and scientific publications.
How accurate is the molecule extraction?
Our extraction achieves over 98% accuracy across thousands of structures from real pharmaceutical patents. Each molecule includes a confidence score so you can prioritize review.
Can I edit extracted molecules?
Yes, DO Patent includes an integrated molecular editor that lets you refine and correct structures before export.
What export formats are available?
You can export extracted molecules as SMILES strings, with full provenance tracking back to source pages and figures.
Ready to extract molecules from your PDFs?
Join researchers at leading pharmaceutical companies using DO Patent to accelerate their workflows.