Every early melanoma story is a story about change.

Ask a dermatologist how a melanoma was caught early and you'll almost always hear the same shape of story. The patient noticed something new. The spouse said it looked different. The doctor, seeing the patient for years, remembered — or didn't. Change over time is the strongest signal in the specialty, and today it lives in the least reliable place we could keep it: human memory.

This isn't a knowledge problem. Dermatologists spend years building the pattern recognition to read a lesion. It's a records problem. In the most visual specialty in medicine, the visual baseline usually isn't in the chart. Photographs exist — in folders, in attachments, taken from different angles in different light — but a photograph you can't line up against last year's is a snapshot, not a record.

The academic world has already shown what the right record makes possible. Structured surveillance programs — standardized poses, every visit, compared against priors — have published case after case of new and changing lesions caught between visits, at the stage where treatment is simplest and outcomes are best. The catch: those programs run on imaging installations that cost as much as a house, dedicated rooms, and staff most practices will never have.

We don't think the record should require the room. The devices are already in the clinic. The medical assistant is already rooming the patient. The visit is already happening. What's missing is the system that turns those five minutes into a standardized, aligned, compounding photographic history — and puts it one glance away from the dermatologist without ever getting in her way.

So that's what we're building, with a rule we refuse to break: the doctor detects, the system supports. Constellation measures, aligns, remembers, and surfaces. It does not diagnose. It does not write the chart. It starts quiet and stays quiet until the dermatologist turns the dial. The moment software starts second-guessing a clinician at scale, it stops being help — and we'd rather build the reference layer she actually consults than the alarm she learns to ignore.

If we do this right, in a few years the question won't be why a community practice keeps a longitudinal photographic record of every patient's skin. It will be why anyone ever practiced without one.