Prolytiq.AI
Start free
ONE PLATFORM · EVERY STAGE

Your entire systematic review,in one place.

From research question to publication-ready figures — without jumping between six different tools. Import, screen, extract, analyze, and report on a single auditable platform.

Start freeBook a demo
Grounded in peer-reviewed methodology · Every decision auditable
screening · in progress
ProjectSR-2026-013 · Anti-IL-6 in Long COVID
847 / 2,031
Tocilizumab for post-acute sequelae of SARS-CoV-2 infection: a randomised, double-blind, placebo-controlled trial
Lancet Resp Med · 2025
Include
Meets P/I/C/O. RCT design with eligible population: adults ≥18 with PASC ≥12 wks.
Cytokine profiling in long COVID — observational cohort from a single tertiary centre
J Infect · 2024
Unclear
No comparator described in abstract. Defer to full-text review.
Editorial: the next decade of post-viral syndromes
Nature Medicine · 2024
Exclude
Editorial — not original research. Excluded by protocol.
claude · audit trail enabled
live
The problem

One review shouldn’t need six tools.

Researchers stitch together reference managers, screening apps, extraction spreadsheets and stats software. Every handoff loses context.

Six tools for one review

Reference managers, screening apps, extraction spreadsheets, stats software, writing platforms. Context lost at every handoff.

No single source of truth

Decisions live in different files. Audit trails break across tools. Methodology becomes hard to reproduce.

Weeks lost to plumbing

Exporting CSVs, reformatting data, reconciling versions. Time that should be spent on analysis.

The workflow

Every stage. One platform.

From framing your research question to exporting publication-ready figures. No tool-switching. No broken audit trails.

04 · T&A Screening
SR-2026-013 · Anti-IL-6 in Long COVID
T&A Screening
Active step
T&A Screening
AI-assisted screening grounded in established systematic review methodology. Include / Exclude / Unclear with full justification.
Roadmap

Coming soon

Q3–Q4 2026
Risk of Bias assessment
Integrated RoB 2.0 and ROBINS-I tools.
GRADE assessment
Certainty of evidence evaluation, built in.
PROSPERO protocol builder
Generate and sync your protocol registration.
PRISMA flow diagrams
Auto-generated from your decision trail.
Methodology

Why trust AI for scientific decisions?

Because our implementation is auditable, reproducible, and grounded in published methodology.

Evidence-first

Grounded in peer-reviewed methodology

Screening follows the decision criteria used by leading systematic review frameworks. Every prompt is inspectable and every decision exportable — you can publish the methodology alongside your review.

# screening prompt (excerpt)
"Decide whether the study meets the
 pre-specified eligibility criteria.
 If insufficient information is
 available in the title and abstract,
 mark as UNCLEAR and defer to full-text."
Audit trail

Every decision, exportable.

Token usage, model version, timestamp, justification — logged per decision and exported as structured JSON, ready for peer review.

// screening_result.json
"decision": "include",
"model": "claude",
"tokens": { in: 1284, out: 312 },
"rationale": "Meets P/I/C/O. RCT design..."
Model quality

AI grounded in reliable models

Powered by Anthropic’s Claude, built with careful prompt engineering and model selection for each stage. Cost-optimized without compromising on quality.

Claude by Anthropic
per-stage selection
Volume screeningfast · low cost
Full-text extractionprecise · nuanced
Adjudicationhuman-in-the-loop
Compliance

Data isolation by default.

GDPR & LGPD compliant. Your data is never used for model training. Each project runs in an isolated database context.

GDPR
LGPD
SOC 2*
Pricing

Transparent pricing.

Free to explore. Pay for real usage, no surprises.

Starter
Free

For exploring & teaching reviews.

  • ›Up to 100 references / month
  • ›1 active project
  • ›T&A screening with Claude Haiku
  • ›CSV & RIS export
Start free
Most popular
Professional
Coming soon

For teams shipping reviews regularly.

  • ›Unlimited projects
  • ›Up to 10,000 references / month
  • ›Haiku + Sonnet orchestration
  • ›Team collaboration
  • ›R scripts & RevMan export
  • ›Email support
Coming soon
Enterprise
Contact

For institutions, hospitals, pharma.

  • ›Unlimited volume
  • ›SSO, granular roles
  • ›On-premise deployment
  • ›SLA 99.9%
  • ›Dedicated onboarding
  • ›Custom DPA
Talk to sales
FAQ

Questions, answered.

How does AI screening work?
Each title and abstract is evaluated against your pre-specified PICO criteria using structured prompts grounded in established systematic review methodology. The AI returns Include, Exclude, or Unclear with a brief rationale. All decisions are logged and exportable.
Can I use my existing PubMed export?
Yes. Prolytiq accepts RIS, BibTeX, PubMed XML, and CSV (NBIB and Endnote formats). You can import from multiple databases and the deduplication step merges overlapping records automatically.
Is my data used to train AI models?
No. Your research data is never used for model training. Each project runs in an isolated database context. You can export a full audit trail at any time.
What statistical analyses are supported?
Prolytiq generates R scripts compatible with the meta, metafor, and robvis packages. Pairwise, network, and proportion meta-analyses are supported depending on your project design.
How does pricing work?
The Starter plan is free with a 100 references/month limit. Professional and Enterprise pricing is based on usage. See the pricing section above for details.

Start your next review today.

No credit card. Cancel anytime.

Create accountBook a demo
sla 99.9% · gdpr · lgpd