Patient
Understand the shape of the conversation.
See how a future patient-facing report could organize findings, explain priorities, and prepare useful questions—without diagnosing or prescribing.
See the pattern, not just the numbers
Laboratory results deserve a clearer clinical conversation.
HealthRadar™ is a clinical information system in development. It is designed to organize laboratory findings into meaningful patterns, priorities, and questions—without pretending to replace the clinician who knows the patient.
Illustrative priority viewSynthetic data
1
What needs attention first
Organize results by clinical priority, not simply by their position on a lab sheet.
2
Which findings belong together
Show related results as patterns while preserving the underlying values.
3
What to discuss next
Convert a dense report into focused questions for the next clinical conversation.
Why this problem matters
Would you bet your health on an answer because it sounds right?
What level of error would you accept in decisions about your health? One mistake in 100? Five? Ten?
Today’s generative AI can explain laboratory results in seconds. It’s fast. It’s often free. And it can write with extraordinary confidence. But confidence isn’t accuracy, and eloquence isn’t clinical judgment.
The question isn’t whether AI can produce a convincing answer. The question is what happens when the convincing answer is wrong about the thing that matters most.
HealthRadar is being developed around a different premise: laboratory interpretation should follow defined, testable rules, produce reproducible results, and make its limits clear. Because when your health is involved, “sounds right” isn’t the standard that should matter.
Questions, not estimates. The numbers above ask about tolerance for error; they are not measured AI error rates.
Choose your pathway
One system. Two different jobs.
Patients and health professionals need different levels of detail, different language, and different decisions. HealthRadar is being designed with that separation built in.
Professional
See how the clinical logic works.
Explore the product model, workflow concepts, evidence approach, research program, and opportunities for clinician-in-the-loop pilots.
Why HealthRadar
A lab report is precise. It is not necessarily organized for understanding.
01
Pattern-level synthesis
Related findings can be considered together instead of treated as an unrelated list of high and low values.
02
Priority architecture
The output can distinguish what appears most time-sensitive from what belongs in routine follow-up.
03
Defined clinical content
HealthRadar does not invent a new clinical interpretation from scratch every time. Its clinical messages, boundaries, versions, and handling of unsupported findings are defined, tested, and reviewed.
Evidence and methods
Reliability should be inspectable.
HealthRadar is being developed with clearly defined clinical boundaries, versioned content, repeatable testing, and documented limitations. We publish the work as it develops—including methods, research findings, limitations, and results that challenge our own assumptions.
Explore the evidence
Our public research library includes:
- The HealthRadar Yardstick benchmark of Google, OpenAI, and Anthropic models
- Early consumer research on VisitPrep
- Working papers describing HealthRadar’s rules-based architecture and evaluation methods
- Selected research on laboratory communication and clinical AI
- Evidence is labeled by publication status, peer-review status, and version.
Start with the public demo
See the organizing idea without sharing a real lab report.
The first public demonstration uses synthetic data only. No account, upload, or personal health information is needed.