Accurate tables
CSV/JSON exports with cleaned structure.
Private beta opening soon
Scientific knowledge, structured automatically.
Ontelya turns research PDFs into clean tables, provenance-linked extractions, and exportable datasets — so you stop copying numbers by hand.
No spam. Early access invites in batches.
Provenance
Every value links back to page/cell/caption.
Fast
CPU-first; GPU only when needed.
How it works
1
Upload
Direct-to-storage uploads, fast and secure.
2
Extract
Docling CPU extraction + quality scoring; GPU fallback if needed.
3
Export
Clean tables + provenance-linked fields in CSV/JSON.
What’s coming after beta
Ontelya Datasets
Build versioned datasets across papers.
Ontelya Analyze
Prebuilt Marimo analyses (compare, rank, correlate).
Ontelya Vision
Extract information from plots and figure tables.
Ontelya Graph
Ontology + cross-paper entity linking.
FAQ
Do you store my PDFs?
Yes, in your workspace storage. Retention limits can be plan-based later.
Is GPU required?
No. CPU by default. GPU only for scanned/image-heavy PDFs.
Can I correct wrong values?
Yes. Corrections are stored as overlays with history and provenance (Pro first).
When do you launch?
When extraction + export is reliable for your target paper types.
Ontelya Extract is live
Turn research PDFs into exportable datasets.
Extract clean tables and provenance-linked parameters. Export CSV/JSON in minutes. Corrections and dataset versioning included.
Extraction
Docling-first pipeline with smart fallback.
Provenance
Every value traceable to page/cell/caption.
Exports
Dataset-ready CSV/JSON + versioning.
Pricing (simple, launch-friendly)
Free
€0
- 10 papers / month
- Basic extraction + export
- Community support
Most popular
Plus
€12
- Higher limits
- Better table cleaning
- Dataset builder (soon)
Pro
€29
- Corrections + verification
- Vision extraction credits
- Prebuilt analyses
Docs
Add your docs later. Keep this section for onboarding: upload → extract → export.