Kagi: what a paid search business must prove

Analysis & application

The strategic question is which recurring research task is valuable enough for someone to pay for. Removing ads is a product choice; it does not, by itself, establish willingness to pay, retention or superior results. A useful comparison starts with the customer’s task and an available alternative.

Kagi’s history dates public beta to 2022 and annual revenue exceeding US$1 million to 2023. ^[1] Annual revenue is neither profit nor money raised. The milestone supports a dated commercial observation, not a claim that the business is currently profitable or that every free category can become a subscription.

Its June 2023 SAFE round raised US$670,000 from 42 accredited investors, mostly users. ^[2] Investment by customers is different from payments for subscriptions. “No venture-capital round” and “no external capital” are not interchangeable descriptions. Any bootstrapping lesson needs to identify who supplied cash and what obligations accompanied it.

A January 2025 interview reported over 37,000 members and roughly 750,000 daily searches. ^[3] These are historical figures, not today’s live dashboard. They do not reveal subscriber retention, acquisition costs or which audience pays most reliably. Dividing searches by members would not establish the behaviour of a typical paying customer.

Kagi’s documentation describes its own indexes combined with external provider APIs. ^[4] A differentiated product therefore need not own every upstream component. The business still needs to consider supplier terms, availability and unit costs. Charging subscribers changes incentives; it does not remove these dependencies or prove that competitors cannot respond.

Analysis & application

1. Define one repeated task, its current workaround and who would pay to improve it.
2. Compare useful results, time to an answer and errors against that workaround.
3. Offer an honest paid trial and measure renewal after a full billing period, not just sign-ups.
4. Track revenue separately from funding; include query costs, support and refunds in the economics.
5. Record what would make you change the offer. Treat this worksheet as a proposed test, not Kagi’s measured internal process.