Create research with sources you can verify1
Generate citation-backed narrative reviews for fast synthesis, or run a systematic review workflow that supports PRISMA 2020 reporting — with AI assistance and human decision control.
1. Every claim links back to the paper it came from — open any citation and read the passage yourself.
A source-linked draft in ~15 minutes·Median 40 references per review·Narrative & systematic workflows
A generated review, every claim linked to its source.
Two paths for research synthesis
Fast when you need a narrative brief. Controlled when you need a systematic review. Both built for health and clinical evidence work, with the source behind every claim one click away.
Narrative Literature Review
~15 minutes
Median time to a source-linked first draft
Systematic Review Copilot
Weeks to months
A review project, every decision on the record
Evidence you can inspect
The reason to trust a synthesis is never that it sounds authoritative. It is that you can follow any sentence back to the paper it came from. Here is that path, end to end.
The search returns papers, not recollections
A question runs against academic databases, the PDFs you uploaded, and your Zotero or Mendeley library. What comes back is a ranked set of real records with real identifiers — the draft is written from those, and only those.
Each claim is written against its own sources
Sentences are drafted from the passages retrieved for that section, and carry the citation for the paper they came from. Citations are checked against the retrieved set as the draft is built; ones that don't resolve are removed rather than left to look plausible.
Any citation opens to the paper behind it
Click an inline citation and you get the title, authors, year, and abstract, with a link straight to the source (Fig. 2). This is the part that matters: you are not asked to trust the synthesis, you are given the means to check it.
Then you can question the evidence directly
Ask across the papers you retrieved — compare statistical approaches, surface disagreements, check what a study actually measured — without leaving the review or pasting anything into a separate chat window.
What opens is the source record itself — title, authors, year, and a link to the original — not the model's restatement of it.
What the record shows
INRA helps researchers save time without giving up the ability to check the work — every figure below is measured across completed reviews on the platform.
- Median time to a source-linked draft
- ~15 minMedian time to a source-linked draft
- Median references per completed review
- 40Median references per completed review
- Reviews completed on INRA
- 10,000+Reviews completed on INRA
“This is a great tool I was surprised how easy it was to use.”
“Writing my dissertation was overwhelming INRA did a good job of helping me organize my thoughts and find relevant studies.”
“The linked citations gave me a ton of confidence in the research and made it so much easier to write my paper.”
Researchers at
One plan, both workflows
Unlimited narrative literature reviews and full access to SLR Copilot. Cancel anytime.
Annual billing works out to six months free.
Professional
$10/month, billed yearly
For researchers who need ongoing access.
- Unlimited narrative literature reviews
- Full SLR Copilot access
- Priority email support
- Word, PDF, LaTeX, and review-data export
Cancel anytime.
Teams & Institutions
$7.50/seat/month, billed yearly
Minimum three seats.
- Everything in Professional, per seat, plus:
- Discounted pricing for multiple seats
- Centralised user management and billing
- Dedicated onboarding and support options
Questions before you start
Start here if you are comparing the two workflows or checking what the platform can and cannot do.
Choose a workflow
Choose a Narrative Literature Review when you need a source-linked first draft for topic orientation, a thesis or grant background, or early scoping. INRA retrieves papers, organizes the findings by theme, and gives you a draft to inspect and revise.
Choose SLR Copilot when your team is running a protocol-driven systematic review. It supports citation import, dual screening, full-text review, extraction, risk-of-bias assessment, and exports for PRISMA 2020 reporting. Reviewers make the final decisions.
ChatGPT is a general-purpose assistant. With search or deep research enabled, it can browse the web, work with uploaded files, and produce reports with citations. INRA is built around literature review workflows.
For a narrative review, INRA retrieves academic papers, organizes the evidence, and produces a draft with citations you can open and inspect. SLR Copilot keeps screening, extraction, risk-of-bias work, reviewer conflicts, and review exports in one shared workspace.
ChatGPT can be useful for brainstorming or explanation. Use INRA when you need to keep the evidence, reviewer decisions, and exports together throughout a review. In either tool, check important claims against the original papers.
Evidence and review decisions
INRA drafts from papers retrieved for your research session and links citations back to source records you can inspect. The pipeline checks citations against the retrieved set and removes references it cannot resolve.
This lowers the risk of fabricated references, but it does not guarantee complete coverage or a correct interpretation of every paper. Before using the work in a thesis, manuscript, grant, or policy brief, check important claims, methods, results, and eligibility decisions against the original sources.
Your team sets the review question, criteria, and reviewer rules. You can then import and deduplicate citations, screen titles and abstracts, review full texts, extract study data, assess risk of bias, and export review records.
AI can suggest screening decisions and help with extraction, but reviewers remain responsible for every final decision. Dual screening, conflict resolution, exclusion reasons, and reviewer activity stay on the record.
Narrative reviews currently retrieve records through Semantic Scholar, PubMed Central, and arXiv. INRA may use Unpaywall to locate an open-access copy, and you can add your own PDFs or supported Zotero sources.
SLR Copilot has its own source workflow. It supports citation-file imports, searches through PubMed, OpenAlex, and Semantic Scholar, and DOI or PMID lookup.
No search system includes every paper or full text. Coverage varies by field, publication date, indexing, and access rights. If an important paper is missing, upload the PDF or citation directly.
Files and outputs
Yes. The available inputs and exports depend on the workflow.
- Narrative reviews: Upload PDFs or bring in supported PDF attachments from Zotero or Mendeley. Export the report to Word, PDF, or LaTeX, and export source metadata as BibTeX or CSV.
- SLR Copilot: Import citation files and full-text PDFs. Export PRISMA 2020 flow diagrams and study lists in formats such as RIS, BibTeX, CSV, and Excel.
Yes. Narrative review output supports languages including Arabic, Chinese, French, German, and Spanish. Retrieval remains English-first, so the language of the report and the language coverage of the evidence are not the same thing. Check specialized terminology and translated claims before using the report.
Account and privacy
INRA encrypts data in transit and at rest and uses access controls. Research queries and uploaded content may be processed by service providers needed to run the platform's AI features. We do not sell your personal information.
The Professional plan includes unlimited narrative literature reviews and full SLR Copilot access. Monthly and annual billing are available, and you can cancel at any time.
Teams and Institutions adds multi-seat pricing, central billing, and onboarding and support options. Contact us if you need a plan for a lab, department, or university.
Build your next review workflow
Start with a citation-backed narrative review, or open SLR Copilot to run a systematic review with your team, on the record.