Everlywell now charges $249 in cash for an AI-generated five-year breast cancer risk score built from a patient-supplied mammogram, the company confirmed last week. The service runs on Clairity Breast, an algorithm that received de novo clearance from the U.S. Food and Drug Administration in May 2026.

The direct-to-consumer launch marks the first time an FDA-authorized AI breast cancer risk model has moved outside the radiology clinic and into a retail website. Everlywell has sold at-home lab tests for ten years; the mammogram-based product adds an imaging-input service to that consumer catalog and signals a new commercialization route for cleared medical imaging AI.

What does the $249 fee actually buy?

An eligible patient uploads a recent mammogram through the Everlywell website. Clairity Breast returns a prediction of that patient's five-year breast cancer risk. The transaction is self-pay; no insurer, provider, or health system sits between the consumer and the algorithm. Everlywell runs the offering through the same logistics chain that ships at-home blood-glucose and celiac test kits.

The site greets visitors with the marketing line "Don't wait. Choose to be informed." — copy the company has carried across its home-test catalog since its founding.

How does Clairity Breast differ from a typical mammography AI?

Conventional mammography AI tools flag suspicious lesions during a clinical read, supporting radiologist workflow. Clairity Breast instead produces a forward-looking risk estimate over a five-year window — a category the FDA authorized through its de novo pathway. That classification matters for R&D managers evaluating imaging portfolios: de novo clearance establishes a new device category, not just a 510(k) equivalent.

The shift moves breast imaging AI from a decision-support tool used by specialists into a consumer-facing screening-adjacent product. Vendors who built mammography AI around radiologist assistance now compete with a model aimed at the patient.

What does the FDA's May 2026 authorization cover?

The agency cleared Clairity Breast for breast cancer risk prediction based on mammographic images. The clearance preceded Everlywell's direct-to-consumer channel by roughly five months. STAT+, which first reported the launch in October 2026, has not released the pivotal study details in the public excerpt: sample size, training dataset composition, validation cohort demographics, and calibration metrics remain undisclosed.

For due diligence, R&D teams should pull the de novo summary from the FDA database before treating Clairity Breast as a benchmark for their own submissions.

Why target consumers rather than radiology groups?

The pricing structure — flat cash, no referral, no imaging-center partnership — sidesteps the reimbursement and IT-integration timelines that slow hospital procurement. R&D managers watching imaging AI commercialization will note that Everlywell has chosen a regulatory-to-retail path: obtain FDA clearance, then monetize directly, rather than pursue enterprise sales cycles that can run 12 to 18 months.

The strategy converts a regulated algorithm into a consumer purchase. A patient who might never reach a breast-imaging clinic now encounters the tool through the same checkout flow used for celiac or cholesterol panels.

What should researchers and R&D managers watch next?

Three metrics will determine whether direct-to-consumer breast cancer risk AI scales: validation performance on mammograms acquired outside the original training sites, uptake among the eligible population Everlywell defines, and downstream clinical action rates — specifically, how many high-risk patients pursue follow-up imaging, biopsy, or genetic counseling.

The October 2026 launch arrives against a backdrop of expanding FDA activity in imaging AI. R&D teams planning 2027 submissions will track whether the agency's de novo framework absorbs the Clairity Breast precedent into its standard review pathway, or treats direct-to-consumer deployment as a separate regulatory question that demands its own evidence package.