Content review
Flags non-standard abbreviations, acronyms undefined at first use, incorrect terms and units, and inconsistent nomenclature across the document.
Features
Everything a reviewer checks — automated.
Medical terminology, statistics, grammar, and consistency — cross-checked against your references and graded on every page.
Built for medical accuracy.
Everything a reviewer checks by hand — terminology, statistics, grammar, and consistency — graded and located on every page.
Statistical and data integrity
Catches percentages reported without an n, mismatched table headers, and figures or conclusions that contradict the data around them.
Grammar, style and branding
Punctuation, spacing, and run-on sentences, plus your style guide — fonts, brand colours, and naming conventions.
References and literature search
Checks claims against the reference pack you upload (PDF, ZIP, or folder) — or searches PubMed and Embase with Boolean terms to find them for you.
Severity-graded issue matrix
Every finding rated High, Medium, or Low with its exact location and a suggested fix — exported to Excel, PDF, or Word, with customisable severity levels.
Private and always improving
Run on-prem or as SaaS, never used to train models, with a feedback loop that tunes the review to your standards over time.
Every review knows who it is for.
Before it reads a word, MedLens is set up for the document context — so the tone and the checks match the market and the audience. This is what sets MedLens apart.
Country / market
Set the market the document is intended for, so region-specific rules and conventions apply.
Promotional or non-promotional
Classify the content so the right compliance lens is applied.
Intended audience
The review tone and depth adjust to the reader. Patient-facing content is held to simpler, plainer language.
Audiences:
Patients
Healthcare professionals (HCPs)
Doctors
Conference attendees
Marketing / Sales reps
Three ways to review.
Pick the depth of review that fits the document and the moment.
Quality review
The full automated compliance and quality check: terminology, statistics, claims, consistency, and references — every finding graded by severity.
Editorial review
A focused pass on grammar, punctuation, spacing, and your company style guide — fonts, brand colours, and naming conventions.
Proofreading (version compare)
Upload two versions of the same document and MedLens highlights exactly what changed between them.
Trained on your standards, not generic rules.
MedLens can learn the documents your organisation already reviews against, so it checks for what your team actually cares about.
Your SOPs and checklists
Provide your standard operating procedures and review checklists. MedLens learns them and reviews every document against your house standard, not a generic one.
Your brand and style guide
Upload brand, editorial, or formatting guidelines and MedLens flags anything off-brand or off-style: fonts, colours, terminology, tone, and naming.
Your severity matrix and outputs
Customise the risk and severity levels, and the output format, to match your internal guidelines.
Upload, run, done.
No integrations, no retraining. Upload a document with its references, and get a graded review back.
-
Upload and configure
Add a PDF, PowerPoint, Word file, or image of any size. Set the country, content type, and audience, and choose the full document or only specific pages.
-
Pick a review mode
Quality for the full compliance check, Editorial for grammar and style, or Proofreading to compare two versions.
-
Add references
Upload references as PDF, ZIP, or a folder — or search PubMed and Embase to pull them in automatically.
-
Run and export
Get a page-by-page issue matrix with severity ratings and suggested fixes, in Excel, PDF, or Word.
Built to catch what manual review misses.
Six issue categories, three severity levels, every page — on every document you upload.
0
Issue categories
0
Severity levels
0
Input formats
0
%
Page coverage
How MedLens compares.
What a careful reviewer does by hand, on every page, in minutes — and the same way every time.
| Capability | MedLens | Manual review |
|---|---|---|
| Turnaround per document | Minutes | Hours to days |
| Page and table coverage | Every page, every table | Sampling, fatigue-prone |
| Severity grading | Automatic High/Med/Low | Reviewer-dependent |
| Reference cross-check | Against your uploaded pack | Manual lookup |
| Output | Excel, PDF, or Word | Tracked changes |
| Consistency across reviewers | Same engine every time | Varies by person |
| Deployment | SaaS or on-premise | Not applicable |
| Data privacy | DPDP and HIPAA-ready | Varies |
Benchmarked, like a frontier model.
MedLens is evaluated on a held-out benchmark of expert-reviewed documents, scored against the same kinds of metrics used to evaluate frontier AI models.
86.7%
Overall benchmark accuracy
Precision
How many flagged issues are real.
Recall (catch rate)
How many real issues it catches.
F1 score
Balance of precision and recall.
Expert agreement
How often it agrees with senior reviewers.
Benchmark figures reflect performance on a representative internal test set and vary by document type and complexity. MedLens output is AI-generated and is intended to assist, not replace, qualified human review, which remains required. Findings should be verified by a competent reviewer before any decision is made. Kredo Analytics accepts no liability for any loss, damage, or misinterpretation of data arising from use of the output.
See MedLens on your own documents.
Book a 30-minute demo and watch MedLens review a document you actually care about.