Stop extracting data by hand.

Oryn screens and extracts data from medical papers in minutes — PRISMA-compliant, learns from your corrections, and scales to any systematic review.

No credit card. No setup. Just early access.

Screening

4,000 abstracts. 40 minutes. Rationale for every single one.

Apply your inclusion and exclusion criteria once. Oryn screens every abstract against them and writes a decision rationale for each one — so you're reviewing reasoning, not reading thousands of abstracts yourself.

Full-text screening when abstracts aren't enough. Liberal inclusion at the abstract stage so nothing relevant gets missed. And it flows directly into extraction — no exporting CSVs to another tool.

4,000+ abstracts in 40 minutes

Validated on a real systematic review

Decision rationale for every abstract

Not a black box — review the reasoning

Abstract → Full-text → Extraction

One continuous flow, no tool-switching

Extraction

AI that extracts like a trained researcher — on every paper, every time.

Paper #200 gets the exact same precision as Paper #1. No fatigue. No drift. No degradation.

Oryn uses a 5-layer prompt system — combining codebook rules, variable-specific instructions, learned examples, and the full paper context — to extract each variable with the same care your best reviewer would on their best day.

Table-aware extraction that reads complex tables, not just running text. Section detection — it knows Methods from Results from Discussion. Every extracted value comes with a confidence score and the exact source text it was pulled from.

5-Layer Prompt System

System → codebook → variable rules → learned examples → paper content

Stateless Architecture

Fresh context for every paper — no quality degradation over time

Table-Aware Extraction

Reads data from complex tables, not just prose paragraphs

Section Detection

Methods, Results, Discussion — targeted extraction from the right section

Confidence Scores

0–100% confidence on every extracted value, so you know what to check

Source Citations

The exact passage behind every number — fully auditable

Same paper. Same variables.

Real extraction from a Phase III oncology trial

Variable
Manual~6 hours
Oryn~3 min
Sample Size
43??
43
98%
Mean Age
62.3
62.3
99%
Median OS
12.4 → 14.2?
14.2 mo
95%
HR (OS)
0.78
92%
95% CI
need recalc
0.61–0.99
94%
p-value
0.03
0.03
99%
ORR (%)
32%? 34%?
34%
96%

3 errors · 1 missing value

All values with source citations

Codebook

1,300+ variables. 30+ domains. Your codebook is already built.

From oncology and cardiology to epidemiology and health economics — every domain includes ready-to-use variable sets with extraction prompts and edge case handling. 66 starter kit templates for common study types get you extracting in minutes.

Need something custom? Build your own variables with types, extraction prompts, validation rules, and conditional logic. If Variable A = X, then Variable B = Y. The kind of codebook rules that used to live in a 40-page manual.

1,300+

Pre-built variables

66

Starter kit templates

30+

Medical domains

If/Then

Conditional logic

Learning

Teach it once. It remembers forever.

Every correction you make is stored with its full context — the paper text, the variable, what went wrong, and the right answer. Next time a similar paper comes through, Oryn retrieves relevant corrections and applies them automatically.

The more you use it, the smarter it gets. Per-project learning by default. Cross-project learning when you opt in. And when the same correction keeps recurring, Oryn suggests new rules so you don't have to keep fixing the same thing.

Corrections stored with full context

Paper text, variable, error type, and the correct answer — all preserved.

Similar papers get automatic corrections

Relevant corrections retrieved as few-shot examples for new extractions.

Per-project and cross-project learning

Learning scoped to your project by default. Opt in to share across projects.

Pattern detection suggests new rules

Recurring corrections become suggested rules — stop fixing the same thing twice.

PRISMA

PRISMA compliance without the paperwork.

Every step follows PRISMA guidelines — from screening criteria to extraction audit trails to decision logging. Compliance is built into the workflow, not bolted on after the fact.

Oryn generates PRISMA flow diagrams from your actual screening data. Every decision is logged: who reviewed, what was decided, when. Export in formats ready for journal submission.

Automatic audit trails

Every screening and extraction decision logged with timestamps and reviewer info.

PRISMA flow diagrams

Generated from your actual data — not a template you fill in manually.

Journal-ready exports

Documentation formatted for submission — no reformatting required.

IPD ModeOnly in Oryn

Patient-by-patient extraction. No other tool offers this.

Auto-detect multiple patients per paper. Extract all variables for Patient 1, then Patient 2, then Patient 3 — structured, clean, ready for IPD meta-analysis.

Case series. IPD studies. The kinds of papers that make extractors groan. Oryn handles them natively.

We looked at what's out there. Nothing else does this.

FeatureOrynCovidenceRayyanElicit
Data extractionPartial
AI screening with rationale
4,000+ abstracts in 40 minManualManualN/A
Screening → Extraction integrated
1,300+ pre-built variables
Learns from corrections
PRISMA compliance built-inPartialPartial
Confidence scores per variable
IPD extraction mode

We extracted data from 150 papers in a fraction of the time it would have taken manually. The accuracy was remarkable — especially on complex oncology endpoints.

Dr. Shouki Bazarbashi·Head of Medical Oncology, KFSHRC

Stop extracting by hand. Start publishing faster.

Join the waitlist for early access. We're onboarding researchers this spring.

No spam. Just launch updates.