Startup Spotlight: Truli turns supplement-labeling headaches into AI compliance software

Why it matters

Truli is building around a compliance problem its founder encountered firsthand, while its own terms make clear that AI review still leaves brands responsible for regulatory decisions.

Catherine Zhou (@cath_zhou) says her attempt to make an iron gummy showed her how hard it was to get a consumer product's regulatory details right. On October 3rd, the Truli co-founder described the problem in a thread on X and said the company would be "launching next monday" while distributing snacks from brands it works with to San Francisco offices.

Truli had already documented product releases in 2025. Its Label Scanner was announced on October 6th, 2025, and its Regulatory AI Assistant followed on November 17th. The October 2026 post does not explain whether the Monday launch refers to a new product, a campaign or another event. It does put Zhou's founder story back in view: she says the compliance problem began with a product she was trying to build herself.

Zhou described her background in a LinkedIn post: four years of research at Stanford, a Nature publication, clinical and EMT work, and software projects including a Chrome extension she said had thousands of users. She wrote that ingredient research for the iron gummy was manageable, while compliant labels and product claims required working through dense rules and costly outside advice. Those details are her account; the X thread itself says she did machine-learning research in drug discovery at Stanford Medical School and published in Nature at 18.

Truli, which a16z Speedrun lists in its seventh cohort, applies AI to compliance reviews for consumer packaged goods. Its current product site describes reviews of labels, ingredients, claims and marketing content, with monitoring across websites and social platforms. The product pages break that work into separate tools: a label scanner, a regulatory question-and-answer assistant, marketing review and social monitoring. The company says scans can flag issues and provide regulatory citations and suggested fixes.

Truli's focus extends beyond packaging on a store shelf. A food or supplement brand can make claims on its label, website, advertisements and social posts. The company says those materials can be checked together, instead of being reviewed only before printing. Zhou says she encountered the problem while trying to sell a supplement, then built software to help other brands handle the same work.

Truli's terms also define the limits of that pitch. The terms of service, last updated September 2nd, 2026, say AI outputs may contain errors and that the service is not legal or regulatory advice. Customers remain responsible for compliance and are told to consult qualified professionals. A missed or misread claim can affect what a brand puts on a package or publishes in an advertisement. Automated review can organize the work and surface possible problems; the warnings in Truli's terms stop short of presenting the software as a substitute for expert judgment.

The post says Truli is backed by Speedrun, and the accelerator's company page identifies Zhou as co-founder and CEO and Michael Wu (@michael_wu8) as co-founder and COO. The X announcement gives no funding amount or valuation. The a16z directory describes Truli as founded in 2025 and lists two employees. Those figures offer a snapshot of the team but do not establish revenue, customer count or adoption. Truli's site says it serves food, beverage and supplement brands; it does not publish customer names or pricing on the pages reviewed here.

Zhou's post also mentions distributing snacks from participating brands. The thread does not identify those brands or connect the sampling plan to a particular product release. Truli's product history shows its approach: use AI for routine regulatory review of products and marketing, while leaving final responsibility with the brand and its advisers.