Ask a clinical question in plain language. EvidenceRN searches PubMed, screens out retracted research, labels every source by level of evidence, and writes a short answer where each claim links to its source.
A project of AI + Nurse Academy. Free and open source under the Apache 2.0 license.
Interface preview. Answer and source text appear here as placeholders and come from live PubMed results in the app.
Each search runs the same four steps, and every step is visible to the person asking.
Converts plain language into a PubMed search and shows the exact query it ran, the filters applied, and the search date.
Retrieves a wide pool of studies, removes retracted records, and ranks the rest by level of evidence, with systematic reviews and guidelines first.
Produces a bottom line, a summary of the evidence, and practice implications within RN scope. Every claim carries a numbered citation.
Flags citation numbers outside the source list and sentences without a citation, so the reader knows where to verify.
Nurses deserve to see how an evidence tool reaches its answers. These features exist so every answer is open to examination.
A panel under every answer shows the question as interpreted, the database searched, the filters, the exact query, and how many records matched.
Each source carries a level-of-evidence label from the Melnyk and Fineout-Overholt hierarchy, inferred from its PubMed publication type.
Retracted publications and retraction notices are excluded before ranking. Expressions of concern, errata, and updates are flagged on the source.
Pulls open-access full text from PubMed Central for the top Level I sources, revises the answer, and lists what the full text added.
Follow-ups draw on the same sources and keep the same citation numbers, so the thread stays traceable.
An urgent-situation banner, RN scope-of-practice language, and a notice that no institutional policy is connected appear on every relevant answer.
EvidenceRN uses public literature, a visible search strategy, labeled evidence levels, retraction screening, and a citation check on every answer. Releasing the code under Apache 2.0 lets the nursing profession examine those choices, test them against real bedside questions, and improve them.
The full application is one readable HTML file and a small server. Anyone with basic technical skills runs it on a laptop in a few minutes.
EvidenceRN is a project of AI + Nurse Academy, which offers nurse-led AI education for nurses and other clinicians.
The project is looking for two or three engineering maintainers, along with clinicians and educators willing to test and review answers. Maintainership comes through contribution, starting with issues and pull requests.
Software developers, health IT, and informatics engineers
Nurses, nurse informaticists, clinical librarians, and EBP faculty
Questions, partnership inquiries, or interest in a maintainer role: hello@evidencern.com
EvidenceRN runs on any Mac, Windows, or Linux computer with Node.js 18 or newer. Search works with no account. The written answers use an Anthropic API key, billed to your own account at a few cents per question.
# download the project git clone https://github.com/EvidenceRN/evidencern.git cd evidencern # add your API key to a private settings file cp .env.example .env # start the app, then open http://localhost:3000 node server.js