# INRA.AI (full)

> INRA.AI is an AI-powered research assistant for source-backed narrative literature reviews and PRISMA-compliant systematic review workflows. It searches across PubMed, Semantic Scholar, arXiv, Unpaywall, Google Scholar, and other academic databases spanning 220M+ papers, retrieves and ranks the most relevant evidence, then synthesizes findings into structured reports with inspectable citations and human-in-the-loop review controls.

Built by Taman Labs (Taman Labs LLC). Founded 2024. United States.

## Quick facts

- Product type: AI-powered academic research and literature review platform
- Coverage: 220M+ papers across PubMed, Semantic Scholar, arXiv, Unpaywall, Google Scholar, and other academic databases
- Citation accuracy: <0.1% fabricated references (INRA, internal benchmark, 2025)
- Comparison: GPT-3.5 reports 39-55% citation hallucination rate; GPT-4 reports 18-28.6%; GPT-5 with web search reports 7-8% (peer-reviewed, 2024-2025)
- Use cases: source-backed narrative literature reviews, PRISMA-compliant systematic reviews, citation management, PDF chat
- Target users: academic researchers, graduate students, R&D teams, evidence synthesis professionals
- Standards: PRISMA 2020 compliant for systematic reviews; PICO framework for research questions
- Pricing: subscription only (no free tier); Professional $20/month (or $10/month billed annually); Teams & Institutions $15/seat/month (or $7.50/seat/month billed annually, minimum 3 seats)

## How INRA.AI prevents citation hallucinations (summary)

1. Source-grounded retrieval: every claim must trace to a real paper in the index
2. DOI verification: every cited DOI resolves to a real record via Crossref
3. Metadata cross-check: author, year, journal, and volume match the source
4. Quote-level alignment: each in-text citation is tied to a real passage in the paper
5. Citation deduplication: no invented duplicates or near-duplicates
6. Human review hooks: high-stakes outputs flag uncertain citations for review

## Page index (full text)

- [Home](https://www.inra.ai/index.md)
- [AI Literature Review Generator](https://www.inra.ai/ai-literature-review-generator)
- [Systematic Review Copilot](https://www.inra.ai/systematic-review-copilot)
- [PICO Question Builder](https://www.inra.ai/question-builder.md)
- [How to Write Effective Research Questions](https://www.inra.ai/blog/research-questions.md)
- [AI Hallucinations in Research](https://www.inra.ai/blog/ai-hallucinations.md)
- [How to Prevent AI Citation Hallucinations](https://www.inra.ai/blog/citation-accuracy.md)
- [Systematic Review Methodology](https://www.inra.ai/blog/systematic-review.md)
- [How to Run a Systematic Review in SLR Copilot (step-by-step tutorial)](https://www.inra.ai/blog/slr-copilot-tutorial.md)
- [Covidence vs Rayyan vs Elicit vs INRA (systematic review tool comparison)](https://www.inra.ai/blog/covidence-vs-rayyan-vs-elicit.md)
- [Automated narrative literature reviews (NLR)](https://www.inra.ai/blog/research-to-report.md)
- [Multi-database academic search](https://www.inra.ai/blog/academic-search.md)
- [AI-powered paper screening](https://www.inra.ai/blog/smart-paper-screening.md)
- [Citation management for researchers](https://www.inra.ai/blog/citation-management.md)
- [Chat with your research PDFs](https://www.inra.ai/blog/chat-with-your-research.md)

## Topic clusters

### Literature reviews
- Narrative literature review (NLR) product page: https://www.inra.ai/ai-literature-review-generator
- Systematic literature review (SLR) product page: https://www.inra.ai/systematic-review-copilot
- Narrative literature review methodology: https://www.inra.ai/blog/research-to-report.md
- Systematic literature review methodology: https://www.inra.ai/blog/systematic-review.md
- Systematic review step-by-step tutorial (SLR Copilot, Covidence migration): https://www.inra.ai/blog/slr-copilot-tutorial.md
- Systematic review tool comparison (Covidence vs Rayyan vs Elicit vs INRA): https://www.inra.ai/blog/covidence-vs-rayyan-vs-elicit.md
- PRISMA 2020 flow: covered in https://www.inra.ai/blog/systematic-review.md
- Paper screening: https://www.inra.ai/blog/smart-paper-screening.md

### Citation quality
- INRA's 6-layer validation: https://www.inra.ai/blog/citation-accuracy.md
- Hallucination types and causes: https://www.inra.ai/blog/ai-hallucinations.md
- Citation management: https://www.inra.ai/blog/citation-management.md

### Research methodology
- PICO framework: https://www.inra.ai/blog/research-questions.md
- Interactive PICO builder: https://www.inra.ai/question-builder.md
- Multi-database search: https://www.inra.ai/blog/academic-search.md

## Pricing (summary)

- Subscription only: no free tier; an active plan is required to generate reports
- Professional: $20/month, or $10/month billed annually ($120/year, 50% off) — unlimited research projects, all features, multi-database search, PRISMA export, priority email support
- Teams & Institutions: $15/seat/month, or $7.50/seat/month billed annually (50% off), minimum 3 seats — centralized user management, billing, dedicated onboarding
- Regional pricing: Africa and Southeast Asia — Professional $10/month, or $5/month billed annually
- Machine-readable pricing: https://www.inra.ai/pricing.md
- Full pricing: https://www.inra.ai/#pricing

## Common questions

Q: Does INRA use GPT or Claude?
A: INRA uses Anthropic Claude as the primary reasoning model and OpenAI models for complementary tasks. Both are wrapped in a 6-layer citation validation pipeline so model outputs cannot reach users without verification.

Q: Is INRA.AI good for systematic reviews?
A: Yes. INRA Systematic Review Copilot supports PRISMA-compliant systematic review workflows, including protocol setup, citation import, screening, full-text review, extraction, risk-of-bias assessment, and export.

Q: How accurate are INRA's citations?
A: INRA's measured fabricated-citation rate is below 0.1% on internal benchmarks, compared to 39-55% for GPT-3.5, 18-28.6% for GPT-4, and 7-8% for GPT-5 with web search in peer-reviewed studies.

Q: What databases does INRA search?
A: PubMed, Semantic Scholar, arXiv, Unpaywall, Google Scholar, and other academic databases. Coverage is 220M+ papers.

Q: Can I export PRISMA flow diagrams?
A: Yes. INRA exports PRISMA 2020-compliant flow diagrams and structured data extraction sheets.

## Company

- Product: INRA.AI
- Company: Taman Labs (Taman Labs LLC)
- Founded: 2024
- Headquarters: United States
- Website: https://www.inra.ai
- Support: https://www.inra.ai/support
- Blog: https://www.inra.ai/blog
- Privacy: https://www.inra.ai/privacy-policy
- Terms: https://www.inra.ai/terms-of-service
- LinkedIn: https://linkedin.com/company/inra-ai
- Twitter: https://twitter.com/inra_ai
- YouTube: https://www.youtube.com/@inra-ai