The Story
Firi has raised $3 million, about โน28.8 crore, in a seed round led by 360 ONE Asset, with Better Capital, CRED founder Kunal Shah and other founders and operators participating.
The Gurugram company sells beauty and personal care products through a curated marketplace with rapid delivery. It uses AI to analyse Reddit discussions, YouTube reviews, ingredient lists and other product information, looking for items with high repeat-purchase potential rather than simply high ratings. It says it has assessed more than 50 million reviews to shortlist around 3,000 products from over 300,000 available globally.
Vivek Madani and Karishma Rathaur incorporated the company in June 2026, after a pilot earlier in the year, and began full operations this month. Both were part of Uber's early India team, where Rathaur worked on categories including Uber Premier and Rentals and Madani later moved to deeptech products at Flipkart Labs. Before Firi, the pair built Honestly, an AI-powered beauty search and discovery platform.
Firi currently stocks skincare, haircare and body care, delivering in as little as nine minutes in parts of Gurugram through what it calls its first Beauty Studio. The fresh capital funds expansion into makeup, fragrances and other categories, along with more dark stores in Gurugram before it enters additional cities.
The company says it is targeting consumers who do not spend much time researching beauty products or consuming beauty content, and who rely instead on recommendations from friends and family.
Why It Matters
Almost every beauty platform in India is built for the person who enjoys beauty. Firi is built for the person who does not.
That sounds like a small distinction and it is the whole company. Nykaa, Amazon, Tira and the rest compete on selection: more brands, more shades, more of everything, on the reasonable assumption that a shopper who cares will want the widest possible choice. It works well for the customer who reads ingredient lists, follows dermatologists on Instagram and knows the difference between niacinamide and salicylic acid. For everyone else, and everyone else is the larger group, three hundred thousand products is not a benefit. It is a wall.
That customer exists in enormous numbers and is poorly served. She wants a face wash that works, does not want to spend two evenings researching one, and currently solves the problem by asking a friend or buying whatever her mother bought. Firi's proposition is to be that friend: a shortlist of around three thousand products, chosen for her, arriving in minutes.
Filtering for repeat-purchase potential rather than ratings is the right instinct behind it. A five-star review tells you someone liked a product enough to write about it once, which in beauty often means they liked the packaging or the first application. Buying the same thing a third time is the only review that cannot be gamed, and it is the closest proxy available for whether something actually works.
The delivery speed matters less than it appears, and matters in a specific way. Nobody has a beauty emergency. What nine-minute delivery does is remove the last hesitation from an impulse: you were told about a serum, you want it before you forget, and waiting three days gives you time to reconsider or to buy it somewhere cheaper. For a company whose entire pitch is removing friction from a decision people avoid, closing the gap between recommendation and possession is the point.
The Strategic Read
The detail that explains this round best is the company the founders built before it.
Madani and Rathaur previously ran Honestly, an AI beauty search and discovery platform. Firi is the same customer problem with the transaction attached. That progression is instructive, because beauty discovery is a notoriously difficult thing to monetise: you do the work of informing a purchase and somebody else collects the money for it. Affiliate revenue is thin, brands pay for placement which corrupts the recommendation, and the user leaves your site to buy elsewhere. Owning the inventory and the delivery is the correction, and it is an expensive one. This is a second attempt at the same insight with a much heavier business attached to it.
Which brings up the part nobody has addressed. The hard problem in Indian beauty retail is not curation. It is authorised supply.
Counterfeit cosmetics are a persistent problem in India, and international brands police their distribution tightly precisely because of it. Nykaa's genuine moat was never its interface; it was a decade of brand relationships and distribution agreements that let it sell products customers trusted were real. A shortlist of 3,000 products drawn from the global market is only as valuable as the ability to source all 3,000 legitimately and keep them in stock, and a meaningful share of those will be international brands with no Indian distributor and no reason to prioritise a one-store company in Gurugram. The AI identifies what to sell. Getting the right to sell it is a different discipline entirely, and it is the one the founders' backgrounds say least about.
The frequency question is the other structural one. Grocery quick commerce works because households order several times a week, which spreads a dark store's fixed cost across enough baskets. Beauty is a monthly purchase at best for most people. Fewer orders per store means the model has to be carried by margin rather than volume, and beauty does have the margin: gross margins in personal care run far above grocery. So the arithmetic can work, but it works differently, and it depends on average order value and repeat rate rather than order frequency. Neither figure has been disclosed, and they are the two that decide this.
Then the signal itself. Reviews are a noisy input in this category. Review manipulation is widespread on large marketplaces, and beauty efficacy is unusually individual: what works depends on skin type, climate and chemistry, so a shortlist derived from global discussion may not map cleanly onto Indian conditions. Filtering for repeat-purchase potential rather than star ratings is a better choice than most, because a product someone buys three times has passed a test no review can fake. Whether 50 million reviews of largely Western consumers predict what works in Delhi humidity is a question the company will answer in its return rates.
What is genuinely in Firi's favour is that it launched full operations this month. Almost nothing about this business has been tested at scale yet, which makes $3 million a reasonable amount to find out with.
For daily, sharp analysis of the biggest moves in the Indian business and startup ecosystem, follow StartupFox.
