THE PAPER TRAIL
The receipts are here: company pages, archived promises, study records and reporting. They establish what Clair said. A sales claim does not become a validation result because it appears in a source list.
Clair’s current pitch. The reservation form promises January 2027 shipping. Exact wording and retrieval metadata were preserved on October 16; the live page can change.
Clair claims training on clinical records, comparisons with at-home tests, and production-device studies. Its clinical-evidence answer sends readers to a Google Forms research-interest intake. These are company claims, not independent validation.
Clair’s own terms define the product as general wellness and limit medical, contraception and fertility-planning uses.
Clair lists its team and advisers. Their titles establish neither institutional endorsement nor product performance.
The supplied 14-page company FAQ records Clair’s prototype results, claimed validation methods, acknowledged limitations and plans for future studies.
Clair’s supplied launch essay reports prototype results, then says the sensor stack changed. This is company marketing, not a peer-reviewed study.
The supplied PDF is titled “Your Body, Your Data,” despite its filename, “Clair - Pioneering Continuous Hormone Intelligence.pdf.” It records Clair’s privacy architecture claims. It is not a security audit.
Mayo confirms program participation. That confirms a place in a program, not completed validation of Clair’s wearable.
The investigator’s unpublished working draft contains opinions, allegations and unfinished notes. It supplies no independent corroboration.
Unpublished author-supplied draft. No public download is available.
Ryan Kern’s article, read in full from the author’s text copy. The investigation separates its allegations from the company, study-registry and patent records checked for this edition.
Full article read from the author’s text copy. The live Substack page remains inaccessible in this review. The supplied text omitted referenced screenshots and several underlying links.
Ryan Kern’s external author link. The linked text could not be read during this review; no findings rely on its contents.
The linked text could not be read during this review. Included only as an external author link.
Shilaih et al. studied cycle-related pulse changes in 91 women across 274 ovulatory cycles. The Scientific Reports paper does not validate Clair’s device.
Speedrun reports $800K in D2C revenue and $18M in B2B contracts signed in under eight weeks. It supplies neither the contracts nor the revenue accounting behind those figures.
Challenges how the B2B agreements are described. The allegation that they are withdrawable letters of intent still needs supporting documents. This investigation does not present it as proven.
Reach’s embedded announcement describes more than 100 fertility-clinic letters of intent. It neither discloses their terms nor establishes that these are the agreements speedrun calls $18M in contracts.
KOS names Agarwal as Head of Machine Learning and promotes Argus performance and a regulatory target. The company issued this announcement; it is not an independent clinical study.
The official registry API records the sponsor’s study plan. Registration is not completed validation. The displayed recruitment status may be outdated.
Verified through the official ClinicalTrials.gov API. The study page may require JavaScript.
The granted claim covers ornamental design. It proves no glucose-monitoring performance.
Google Patents records abandonment on October 7, 2025 for failure to respond. Its legal-status disclaimer applies. This record does not settle the full portfolio or prosecution history.
Agarwal identifies himself as Clair’s CTO and co-founder and describes bringing his earlier glucose-wearable experience to Clair. His account supplies no independent product validation.
The clinical-evidence FAQ links to this Google Form. It asks for prospective research interest, contact details and health-background information.
Public introduction and question definitions reviewed. No responses entered or submitted. The form can change.
UW permits institutional Google Forms for some research surveys, with restrictions. Platform choice alone proves neither legitimacy nor fraud. UW’s rules do not govern Clair or establish its approval.
All 13 supplied images were visually reviewed and preserved unchanged in the deck reader. The public deck is redacted. It does not show everything investors received.
All 13 supplied images reviewed. Blacked-out numbers, names and terms remain unknown.
Geoff Weiss identifies this as the deck used for Clair’s $11.6 million seed raise and says it was redacted for public sharing.
Clair labels these “B2B device purchase MOUs.” The count and represented value are blacked out. The slide gives late 2026 as the planned shipping date.
Clair pitches a clinical-partner network ahead of a November launch, then labels the list “OUR LOIS INCLUDE.” The partner count and list are blacked out.
Clair claims a 23,000+ waitlist and 5,600+ preorders beside a rising cumulative-preorder chart. The represented monetary value is blacked out.
Clair promises a 10-biosensor stack tracking 130+ biomarkers without blood, urine or saliva. The pitch says a sensor-fusion model turns those signals into continuous hormone readings.
The app screens display estrogen at 268 pg/mL. Underneath, Clair claims validation against gold-standard hormone assays in a real-world clinical trial.
Source for the timeline’s short shipping-date excerpt.
The supplied homepage image promises December 2026 shipping for Founding Members and is preserved byte-for-byte. Its filename supplies the stated date; no independent archive timestamp verifies it.
Ron Leuty reports a planned Q3 formal launch. February reporting already named November: the targets conflict, but this does not establish when Clair revised them. Leuty also reports that Duan paused her sociology master’s studies to build Clair.
Reports a November hardware launch before the March report of a Q3 target. This preserves the conflicting chronology. It does not validate the article’s technical claims.
Clair’s transaction counsel names the announced round, lead and seven participants. It discloses no individual check sizes or ownership.
Reach names itself and a16z speedrun as pre-seed co-leads. The reviewed newsletter gives no round amount.
Jomayra Herrera explains Reach’s investment rationale and describes Agarwal’s earlier glucose-monitoring work. This is the investor’s account, not independent validation.
A Treehub partner describes its first check, demand-building work and later financings. This is an investor’s account. It does not open the private diligence file.
Treehub names Clair among its initial investments and identifies its relationship with AI Health Fund.
Lists Roxana Daneshjou on the team and links to Treehub. It discloses no personal Clair equity, compensation or role in a validation experiment.
Morgan endorses the founder’s funding announcement. The post supplies no technical validation.
Confirms Morgan’s partner role and scientific and medical training. Those credentials do not validate Clair’s product.
Forerunner says it led Oura’s December 2019 Series B. That establishes Oura backing. It says nothing about a decision on Clair.
Dexcom announces a $75 million strategic investment in Oura’s Series D. It does not say whether Dexcom considered or declined Clair.
Names Abbott as a strategic investor in WHOOP’s Series G. It does not disclose Abbott’s full holdings or a decision on Clair.
Duan’s interview biography identifies her completed Stanford B.S. in Symbolic Systems. Her product-performance statements remain company claims, not independent validation.
Stanford lists Duan as a Sociology M.A. student admitted Autumn 2023. Enrollment does not establish that she earned the degree.
Stanford CS’s embedded post identifies Agarwal as a Computer Science alumnus with B.S. and M.S. degrees, and Duan as a B.S. alumna. It verifies neither product performance nor a personal relationship.
Lists Abhinav Agarwal as a Computer Science B.S. and M.S. degree candidate on printed pages 14 and 25. Candidacy alone does not confirm conferral; Stanford CS’s later alumni post supplies further evidence. The profiles omit graduation years because the records differ in timing.
The supplied poster photograph names mcPHASES, Fitbit Sense, Dexcom G6 and Mira in its methods. It reports anovulation detection and a separate Kerala PCOS analysis. The event and affiliations are the poster’s own statements; it shows no exact presentation date. Its model results were not independently reproduced.
Methods and provenance for the dataset named on Clair’s poster. DOI 10.13026/zx6a-2c81. Participant files require registered access and a data-use agreement. This review did not access them.
The dataset paper credits Georgianna Lin, Jin Yi Li, Kaavya Kalani, Khai N. Truong and Alex Mariakakis, all at the University of Toronto. It reports collection methods and 192 complete cycles under its stated definition. DOI 10.1038/s41597-026-06805-3.
The state investment summary dates the firm’s founding to 2004.
Describes $1M–$2M seed investments at that time. This is historical strategy. It supplies neither a current fixed range nor Clair’s check size.
Dates speedrun’s launch to 2023. The program’s own alumni outcomes are the relevant record for the comparison with Y Combinator.
Publishes $500K for 10% upfront through a SAFE, plus another $500K in the next round within 18 months. These program terms do not disclose Clair’s actual deal.
Dates Reach’s independent founding to 2015 and describes its earlier NewSchools Seed history. Publishes a typical range from $100K at pre-seed to upwards of $12M at later stages.
Announces Treehub’s launch and first spring cohort. Names Anne Wojcicki as Operating Partner and Esther Wojcicki as Founding Advisor at launch.
Treehub’s event listing advertises up to $150K for selected teams. That is a program ceiling. It does not disclose Clair’s check.
The organizer’s event listing proves Brydge Club was active by this date. It does not establish when the club was founded.
Advertises $1K–$50K checks and lists Clair alongside co-investors a16z and Khosla. A per-check range does not cap members’ combined allocation. Its product language is promotional.
Names Bellis as founder and general partner, dates Cartan’s founding to 2023 and describes her earlier venture-capital experience.
Reports a $500K Fund I average and planned $1M–$2M Fund II checks. Bellis describes co-investing, performing diligence and cross-checking with lead investors. These dated strategy statements disclose no Clair allocation.
Article text retrieved through indexed search. Direct access may require a subscription.
AGI House in Hillsborough, California, dates Rocky Yu’s launch to 2023 and advertises investments up to $1M. This is the organization examined in the funding profile.
Official text retrieved through indexed search. The live page uses client-side rendering.
Licensed New York Times reporting by Kirsten Grind and Minh Connors covers police incidents at the Hillsborough property, a woman’s allegation involving a resident, and Rocky Yu’s responses. The incident count is not a count of sexual-assault cases. The report gives no adjudicated outcome for that allegation.
Names FACEIT founders Niccolo Maisto and Michele Attisani and describes their operating history. The reviewed page gives no check-size range.
Names Insiders Ventures Advisors Limited as the website operator, linking the firm’s site to its UK company record.
The adviser’s UK company record establishes incorporation. It does not necessarily date the fund’s founding or first investment.
Insiders identifies London as its headquarters and describes its follower-investor model. Its funding post confirms participation in Clair’s Khosla-led seed. The location and follower label do not reveal what diligence it performed.
All seven independent directors explain their strategic disagreement and concerns over concentrated voting control in this SEC-filed letter. It makes no finding of personal fraud.
The SEC-filed company announcement confirms Chapter 11 and Wojcicki’s resignation as CEO. It does not establish that she deliberately caused the bankruptcy.
The ICO’s final findings establish company security failures behind the 2023 breach. It says the identified infringements had ceased by the end of 2024. The company received the penalty; this is no personal fraud finding against Wojcicki.
The SEC 8-K records TTAM’s purchase of core businesses for $305M cash plus assumed liabilities, with some assets excluded. Wojcicki did not personally buy every former operation in this transaction.
Names Anne Wojcicki as founder and CEO of the successor research institute. Documents the continuing business after the asset acquisition.
Duan’s funding post tags AGI House with Rocky Yu and Andy Zhong, identifying the Hillsborough organization. It names other participating firms and individuals but discloses no individual check sizes.
YC cofounder Jessica Livingston records its 2007 investment in Dropbox and celebrates its first portfolio IPO. She places YC’s launch 13 years earlier, in 2005. This is a completed outcome; it also shows how long a track record can take to develop.
Identifies Airbnb as a Winter 2009 YC company and lists its status as public. This is a concrete alumni outcome in the comparison with speedrun; it says nothing about the accuracy of Clair’s bracelet.
Greenhouse confirms the acquisition closed and names a16z Speedrun among Ezra’s seed backers. This is a documented speedrun portfolio exit. The release does not disclose the purchase price or investor return, so the comparison does not treat it as a measured financial success or claim speedrun has no outcomes.
The paper concerns structured pruning of a language model. The abstract describes relative performance retention on language benchmarks; Table 2 reports task-specific accuracy. It is not a clinical validation study of a glucose or hormone wearable.
Identifies SynapticRAG as a Stanford CS224W Fall 2024 course project. It is a Medium post in the course publication; the page explicitly discloses this context.
Article text was readable during review. Medium blocks some automated requests and controls visitor access.
Linked from the official archived course project page. All three project types require a blog post; instructions describe course-staff review and subsequent publication on the course Medium page. Coursework is real, but this process is distinct from journal publication.
Names Abhinav Agarwal and links his abhinav30219 account. It establishes a credited contribution, not his total hours or full participation period.
The acknowledgments name Abhinav Agarwal for PAWS-team work. He is acknowledged, not a named author. This is evidence of participation, not validation of Clair’s bracelet.
Credits Abhinav A. among the course students who built the PAWS mobile app, linking his profile. This further corroborates his contribution.
Try a different word or choose “All sources.”
Published sources are linked individually above. External sites control access to their pages. These are the records under examination; listing them does not endorse their claims.