AI Clinical Note Validator

RAG + Rule Engine + Local LLM + Batch Review

Portfolio showcase animation

Enter Clinical Note

Validate Note

Completeness Score

100%

Risk Level

Low

Confidence

High

Overall Status

Good

Validate a clinical note

Section-aware parsing, rule checks, and explainable scoring

Enter Clinical Note

Chief Complaint: Chest pain and shortness of breath. Diagnosis: Acute coronary syndrome. Treatment: Patient received aspirin and nitroglycerin during hospitalization. Medications: Discharge medications include aspirin and atorvastatin. Follow-up: Follow up with cardiologist in 1 week.

Validate Note

Single-note validation

Strong note example with complete sections and low risk

Completeness Score

100%

Risk Points

8

Confidence Score

92%

Overall Status

Good

Recommendations

No major gaps detected

Diagnosis specificity present

Follow-up complete

Detected Sections

Symptoms

Diagnosis

Treatment

Medications

Follow-up

Rule Summary Chart

Passed · 5

Warnings · 0

Failed · 0

Rule engine catches documentation gaps

Weak note example with missing sections and high risk

Completeness Score

52%

Risk Points

71

Confidence Score

41%

Overall Status

High Risk

Recommendations

Add treatment details

Document medications

Add provider + timeline follow-up

Documentation Gaps

Missing medications

Missing follow-up

Weak diagnosis specificity

Rule Summary Chart

Passed · 1

Warnings · 2

Failed · 2

Batch note upload

Validate multiple notes and review summary metrics

Uploaded File Preview

Chief Complaint: Chest pain...
---NOTE---
Patient admitted with chest pain. Diagnosed and discharged.
---NOTE---
Patient presented with fever and cough...

Process Batch

Batch Summary

note_id	overall_status	risk_level	risk_points	completeness_score	confidence_score
NOTE_1	Good	Low	8	100	92
NOTE_2	High Risk	High	71	52	41
NOTE_3	Needs Review	Medium	38	76	68

Validation history and export

Review saved reports and download previous results

Saved Reports

created_at	overall_status	risk_level	score
2026-04-23T10:15:30	Good	Low	100%
2026-04-23T10:17:02	High Risk	High	52%
2026-04-23T10:18:41	Needs Review	Medium	76%

Project Highlights

RAG-powered guideline retrieval

Hybrid rule engine + LLM reasoning

Completeness, risk, confidence scoring

Batch upload + history + reports