The pitch changed. The claim did not disappear.

Change the wording. Leave the claim unanswered.

The archived FAQ says Clair already estimates hormone concentrations and describes a future upgrade that would display numerical hormone levels. The current homepage leads with “hormone-aware health.” That is a useful change for the copywriter. It does nothing to establish the concentration-estimation capability Clair had already advertised.

Clair should state whether it stands behind that capability, publish the testing, and explain the change. A customer should not need to excavate an old FAQ to work out what the bracelet is supposed to estimate. The old claim needs an answer, not a softer adjective.

The archive preserves an essay dated January 23 and a FAQ dated February 6; the FAQ copy carries an October 8 print timestamp. I recorded a website refresh that evening. These are records of wording, not a deletion log or proof of why a page changed. They preserve the specific promise. A new page does not erase it.

Announcing a study is the easy part.

The January essay announced an independent study through the Stanford Gladstone BeeHive program starting in May 2026. The later-dated FAQ laid out three phases: data collection beginning in April, refinement and robustness work, and independent validation involving 150 women. A peer-reviewed publication also made the roadmap. The announcements were ambitious. The follow-through now needs documentation.

Publish the study registration or protocol, the current recruitment status, the hardware used, and the results for completed work. These announcements establish plans; they do not establish completion or settle whether the announced study is underway. That requires a current record, not another mention of Stanford.

Clair was specific enough to announce the study. It can be specific enough to report what happened.

Privacy, guaranteed by a paragraph.

The preserved privacy article describes on-device analysis, local storage, optional encrypted backups, and separate choices about sharing data for research. It makes unusually strong claims about Clair’s inability to access health data and attributes that protection to the system’s architecture. Good. Architecture can be examined.

Show the implementation, identify the exceptions, and provide an appropriate security review. If privacy is an architectural guarantee, make the architecture carry it. A description of the intended system does not establish how the production system works. Marketing copy is not a security test.

The archive preserves the promise. It does not verify the production system. Clair has published the assurance; it still needs to provide the evidence that lets readers check it.

November. December. January.

The product pitch keeps its confidence. The calendar keeps moving. TechCrunch reported on June 17 that Clair planned to ship units in November. The homepage screenshot supplied with this investigation, named October 8, promises December shipping for Founding Members. On October 16, the live reservation form says January 2027. These are shipping targets, not the clinical-study schedule.

The timeline puts the dated claims side by side, with the exact language and direct sources. The December screenshot remains unchanged; the January wording carries its retrieval date. Clair can change the live page. It cannot change what these records say.

As of October 16, all three advertised shipping months are still future targets. This is a record of later advertised timing, not three already-missed deadlines or proof that every customer’s order was delayed. The deck’s separate November launch is not a shipment date, and these documents do not identify every order batch.

Publish the original and revised schedules. Identify which orders, if any, received later dates, when customers were told, and why. There is a straightforward way to explain this record. Replacing the month beside the reserve button is not it.

Compare the shipping promises →