Analysis Workbench
Analyze clinical text with AI-powered contextual understanding
AI output is decision support only.
Treat every summary, differential, and recommendation in this workspace as preliminary. A licensed clinician must independently verify the source data, interpretation, and any resulting clinical action.
Document Upload
Click to upload or drag and drop
PDF, TXT, MD, CSV (max 10MB)
Clinical Text
0 charactersAnalysis Options
Analysis Result
Analysis results will appear here.
Paste clinical text and click Analyze to begin.
Evidence and Citations
Inline citations in the analysis map to these indexed note references.
Citation legend: [NOTE] = the original note you pasted, [NOTE-CHUNK-1], [NOTE-CHUNK-2], etc. = retrieved indexed passages shown below.
Watch the AI think
Run an analysis to view the processing trace. This shows retrieval and workflow steps, not hidden model reasoning.
Clinical Review
Multi-turn clinical consultation with structured evidence synthesis
Clinical Review is advisory, not diagnostic.
Use this panel to explore hypotheses and evidence, not to issue patient care decisions. A licensed clinician must confirm the underlying chart facts, evidence relevance, and final clinical judgment.
Imported From Analysis Workbench
This clinical review will include the source note and prior workbench analysis as background context.
Imported Note
Imported Analysis
Suggestion Chips
Suggested questions to explore the value of Clinical Review
Selecting a chip fills the draft only. Press Run Clinical Review to send.
Consultation Thread
Idle
Start a clinical consultation.
Ask questions about clinical scenarios, differential diagnoses, or treatment plans.
External Literature Grounding
PubMed OnClinical Review now always queries PubMed so conclusions can be anchored in external literature where available.
PubMed Query Used
PubMed query generated automatically from the latest review question.
Supporting Note Evidence
Inline [N#] citations map to indexed note passages shown below.
Session Info
- Messages
- 0
- Model
- Pending
Watch the AI think
Run a consultation to view the AI's reasoning process and workflow steps here.
Recruiter Test Kits
Pre-loaded synthetic clinical scenarios for evaluation. Load any case into the Workbench, then analyze or send to Clinical Review.
Tip: Load a recruiter kit, summarize it in Workbench, then move into Clinical Review for higher-level follow-up questions.
Guided Demo
Core Workbench
2 synthetic casesDischarge Summary Review
Medication reconciliation plus follow-up verification use case.
Suggested prompts:
Lab Trend Integrity Check
Flag abnormal values and compare with plan-of-care alignment.
Suggested prompts:
ICU Patients
2 synthetic casesICU Longitudinal Review: Septic Shock and ARDS
Multi-week surgical ICU stay with source control, ARDS, CRRT, tracheostomy, delirium, and competing complications.
Suggested prompts:
Recruiter Scoring Rubric (1โ5)
Scored: 0/4Timeline fidelity: Correct sequence and date/day alignment for source control, ventilator changes, CRRT, tracheostomy, and transfer milestones.
Complication synthesis: Clearly identifies each major complication, cites objective evidence, and states resolved vs unresolved status at transfer.
Clinical reasoning quality: Distinguishes correlation from causation and explains why major management pivots occurred.
Grounding and traceability: Uses specific chart facts/labs/events from the note rather than generic ICU language.
ICU Longitudinal Review: Post-CABG Cardiogenic Shock
Complex cardiothoracic ICU course with VA-ECMO, CRRT, embolic strokes, GI bleeding, failed extubation, and recovery planning.
Suggested prompts:
Recruiter Scoring Rubric (1โ5)
Scored: 0/4Hemodynamic support accuracy: Correctly tracks rationale and timing of IABP, VA-ECMO, inotropes, decannulation, and recovery signals.
Risk balancing analysis: Explains anticoagulation decisions in context of stroke, bleeding, and postoperative instability.
Cross-domain integration: Connects cardiology, neurologic, renal, respiratory, and GI events into a coherent ICU narrative.
Transfer readiness assessment: Uses explicit objective findings to justify why CTICU transfer was appropriate despite residual risk.
FHIR Explorer
Browse and query FHIR R4 resources, assemble a patient picture, and import to the workbench for AI analysis
FHIR Server
Resource Query
Resource Cart
0Add resources from search results
Results
Connect to a FHIR server and search for resources.
Click + on any result to add it to your cart, then Compose & Import.
๐ Composed Clinical Summary
Benchmarks
AI performance test suite and operational metrics
Avg Latency
โ
Total Cost
โ
Avg Consistency
โ
Success Rate
โ
| Test | Avg Latency | Tokens | Cost | Citations | Consistency | Success |
|---|
Total Requests
โ
Avg Latency
โ
Error Rate
โ
Uptime
โ
Route Performance
| Route | Requests | Avg (ms) | P95 (ms) | Errors |
|---|---|---|---|---|
| Click Refresh to load metrics | ||||
Model Usage
No model usage data available