Verify a Claim

Paste your claim and provide the reference. RefCheckr will find the supporting passage and assess the strength of evidence.

⬆
or drag and drop
📄 ✕
or
Parsing reference text…

Analysis

Find References

Search PubMed, ClinicalTrials.gov and drug labelling databases for references that support your claim.

Search across
Generating search strategy…

Results

History

Your recent verifications and searches.

↺

Coming soon

Your verification history will appear here once you start checking claims.

How it works

RefCheckr helps medical writers verify claims against their references with precision.

Verify a Claim

Check whether a specific claim is supported by its cited reference.

1
Paste your claim
The exact sentence or data point you need to verify — including any specific numbers, p-values, or statistical claims.
2
Provide the reference
Upload the source PDF or paste the reference text. RefCheckr parses the document structure — pages, columns, paragraphs, and line numbers.
3
Get your verdict
RefCheckr locates the exact supporting passage, highlights it in context, and delivers a verdict: Strong Support, Partial Support, Not Supported, Contradicted, or Overclaim.

Find References

Search multiple databases for references that might support a claim.

1
Enter a clinical claim
Describe the claim you need to find supporting evidence for. RefCheckr generates an optimised search strategy automatically.
2
Choose your sources
Search across PubMed (literature), ClinicalTrials.gov (trial registrations), and DailyMed/OpenFDA (US prescribing information). Toggle sources on or off.
3
Review ranked results
Results are grouped by source and ranked by relevance. Each PubMed article includes a brief assessment of how it relates to your claim.

What powers RefCheckr

Built on established biomedical infrastructure, not just a language model wrapper.

Analysis
Perplexity AI (Sonar) for claim verification and relevance assessment
Literature
PubMed / NCBI via PubCrawl MCP server
Trials
ClinicalTrials.gov API via PubCrawl
Labelling
DailyMed (FDA structured product labels) and OpenFDA via PubCrawl
PDF parsing
Page, column, paragraph and line-level extraction with structural markers

Frequently asked questions

Common questions about RefCheckr.

Is RefCheckr free?
Beta access is free with an invite code. Paid plans with higher usage limits and team features are coming soon.
What file formats are supported?
PDF and plain text. PDFs are parsed with full structural awareness — pages, columns, paragraphs, and line numbers — so RefCheckr can pinpoint exactly where a supporting passage lives.
Which AI model does RefCheckr use?
Perplexity AI (Sonar) for claim analysis and relevance assessment. Literature search, clinical trials, and drug labelling are handled by PubCrawl, which queries PubMed, ClinicalTrials.gov, and DailyMed directly — no LLM in the loop for those lookups.
Does RefCheckr store my documents?
No. PDFs are processed in memory and never written to disk or stored on our servers. Your documents stay yours.
How accurate is the verification?
RefCheckr is designed to catch overclaims, unsupported statistics, and mismatched data — the kinds of errors that slip through manual checking. But it's a verification aid, not an infallible oracle. Always apply professional editorial judgement, especially for regulatory submissions.
Can I use RefCheckr for regulatory submissions?
RefCheckr supports your reference-checking workflow but is not a validated regulatory tool. Use it for first-pass verification and to flag potential issues — then confirm critical claims manually before submission.

A verification aid, not a replacement for judgement. RefCheckr is designed to support medical writers in their reference-checking workflow. It identifies relevant passages and flags potential issues, but professional editorial judgement should always be applied to the final assessment. Always verify critical claims manually.

Settings

Manage your access and preferences.

Access

Enter your invite code to use RefCheckr.

Contact

Questions, feedback, or partnership enquiries — we'd like to hear from you.

Get in touch

RefCheckr is built by Nick Lamb at PharmaTools.AI. Reach out for bug reports, feature requests, or to discuss enterprise access.

✉
Email
nick@pharmatools.ai
 in
LinkedIn
Nick Lamb
 ↗
Website
pharmatools.ai