Real-world evidence extraction
Build RWE cohorts from electronic health records without months of manual abstraction. Define phenotypes, run extractions, export analysis-ready data.
Learn moreClinical NLP for Evidence Teams
Turn clinical notes into structured, research-ready evidence.
CliniNote reads the documentation your teams already write and extracts structured, real-world evidence you can query, audit, and trust.
Input: discharge summary
Patient presents with Type 2 diabetes mellitus (E11.9) and a history of hypertensive heart disease (I11.9). Currently prescribed metformin 1000mg twice daily and lisinopril 10mg. Follow-up HbA1c panel scheduled for 2025-03-14.
| Entity | Type | Code | Conf |
|---|---|---|---|
| Type 2 diabetes mellitus | Diagnosis | E11.9 | 0.97 |
| Hypertensive heart disease | Diagnosis | I11.9 | 0.94 |
| metformin 1000mg | Medication | NDC 68180-221 | 0.98 |
| HbA1c panel | Procedure | CPT 83036 | 0.88 |
Diagnosis
Medication
Procedure
Date
The chart review problem
Chart review at scale is broken
15-45 min
to manually abstract a single clinical note
60-80%
of study time spent on data extraction, not analysis
12-30%
inter-reviewer disagreement on unstructured diagnoses
RWE teams at hospitals and CROs spend the majority of their time extracting data that already exists in clinical notes, not doing the analysis they were hired to do.
Manual chart review is slow, inconsistent, and impossible to audit. CliniNote replaces the extraction step entirely.
Figures represent industry-observed ranges from published clinical research methodology literature.
How it works
From EHR export to structured dataset in three steps
Connect your EHR export
Upload FHIR bundles, HL7 v2 files, CSV note exports, or paste raw clinical text directly.
Configure what to extract
Select from standard entity types - ICD-10 diagnoses, NDC medications, CPT procedures, lab values, and clinical dates - or define custom phenotypes.
Download your structured dataset
Receive a clean CSV or JSON with extracted entities, confidence scores, and source sentence citations.
Built for clinical research
Extraction that meets the standards of evidence work
Clinical NLP trained on medical text
Purpose-built models fine-tuned on de-identified clinical notes - not general-purpose LLMs that hallucinate medication names and miss negations.
Configurable entity types
Extract exactly what your study protocol needs: diagnoses, medications, procedures, lab results, dates, and free-text findings. Add custom entity definitions via the portal.
Source citations on every row
Every extracted value links back to the exact source sentence, so reviewers can verify and auditors can trace.
Structured audit trail
Each extraction run logs the model version, configuration, confidence thresholds, and reviewer overrides - exportable for regulatory submissions.
We used to spend two weeks extracting exposure data from cardiology notes for a single cohort analysis. CliniNote does it overnight, with citations we can actually verify.
88% F1
on primary diagnosis extraction
Across a 400-note internal benchmark set, measured against double-coded manual review.
Use cases
What clinical teams use CliniNote for
Clinical trial eligibility screening
Screen patient records against protocol inclusion and exclusion criteria. Surface eligible candidates from structured extraction rather than full chart review.
Learn moreOutcomes and safety research
Track adverse events, medication patterns, and procedure outcomes longitudinally across note series. Get structured timelines, not free-text paragraphs.
Learn moreSecurity and data trust
Built with the controls clinical data requires
Your data does not train our models
Patient note data you upload is never used to train or improve CliniNote models. Extraction runs are processed in isolated environments and not retained beyond your session unless you explicitly save.
Designed with HIPAA controls in mind
Data handling follows HIPAA-adjacent principles: encryption at rest and in transit, role-based access, audit logging. We work with institutions to establish appropriate data agreements.
Granular access controls
Role-based permissions, SSO-ready authentication, and per-project data isolation. Admins control who can upload, extract, review, and export.
Transparent pricing
From $349 per month
Starter plan covers up to 3,000 notes per month for small research teams. Research plan scales to 15,000 notes with custom entity types. Free 14-day trial on both.
Starter
$349/mo
Up to 3,000 notes/month
Most popular
Research
$999/mo
Up to 15,000 notes/month
Institution
Custom
Unlimited notes