Capsa AI: Product, team and how it was built

Solo founder · sold 17 months after launch

It solved one narrow and repeated task that every company has at the end of each quarter: find all invoices and give them to the accountant. It read the invoices where they already were, in the inbox and on supplier websites, instead of asking people to forward them, and it let customers write their own checking rules in plain language. Stripe data shows monthly revenue rising from US$12,631 in November 2025 to US$18,970 in January 2026.

Team of 2 to 10 · 33 months after launch, revenue falling

The product goes where the customers already are: a shop owner does not install a new app or learn accounting software, but writes into a chat. The founder wrote of having worked as a data scientist at a microloan company and of running a café, so the founder knew both the bookkeeping problem and the reason banks ask for records.

Solo founder · 19 months after launch, revenue growing

The founder, a digital nomad, had collected more than 1,000 receipts from many countries to track his own spending, and built the app for that use. It reads every line item instead of only the total, which matters for travellers with receipts in foreign scripts and currencies. Small, steady releases (25 versions between 21 January and late October 2026) and a translation of the app into 52 languages widened the market without a team.

Team of 2 to 10 · 24 months after launch

It picked work that large regulated companies had failed to automate for years and that tolerates almost no errors. Instead of letting a model decide each step, it turns the expert's procedure into fixed program steps, which makes the result predictable and cheap to run. A few large customers, sold directly, can pay for a very small team.

Started small, grew after funding · 25 months after launch, revenue growing

Every GmbH and UG in Germany must keep books, file VAT returns and produce annual accounts, and most pay a tax adviser's office to do it, the company's press release says. mika puts AI on the routine part, keeps people on the checking, and sells it as a monthly subscription that the price page lists from €49 net. It chose one narrow customer, the small limited company without its own finance staff, and the customer count grew from about 400 at the start of 2026 to more than 750 in September 2026, the company said.

Team of 2 to 10 · 7 months after launch

Chasing late invoices is repetitive work that finance teams do by hand in spreadsheets and email. Cleavr connects to tools that companies already use, such as Pennylane, Stripe and SAP, so a customer can start without a migration, and the company says setup takes under a day.

Started small, grew after funding · 42 months after launch

Companies in finance and insurance want to automate support but cannot accept an AI that invents a refund policy. The founders saw support at Uber's scale and built the product around that risk: a confidence threshold, an audit trail, and a hand-off to a person. The company also sells a 90-day guarantee, which its home page states as 90% resolution in 90 days or the customer pays nothing.

Team of 2 to 10 · 16 months after launch, revenue growing

Four owners of marketing agencies built it for their own clients, who asked whether AI recommended them. They launched it with plans made for agencies and a 30% lifetime commission for agencies that resell it. The average subscription now pays about US$206 a month, close to the price of the agency plans.

Solo founder · 15 months after launch, revenue falling

It does the work in the cells instead of explaining what to paste, and it works in the Excel that people already have, without a Microsoft Copilot license. The founder already taught Excel to a large French audience. The AI models are bought by the request from OpenAI and Anthropic, so one person could build and run it.

Team of 2 to 10 · 8 months after launch, revenue falling

Filling in the same facts on every employer's application form, and rewriting a resume for every job, is the slowest part of a job search for students and new graduates. The founder had already built a growth system for her first app: hundreds of paid short-video creators trained with her own course. She moved the same system to a job app, aimed at college students and recent graduates, who she says have more buying power than her beauty app's users.

Team of 2 to 10 · 4 months after launch

The founders had worked inside healthcare billing before, so they knew that insurance follow-up calls are a large, repetitive and expensive part of the work. A voice model can now hold this kind of short, scripted business conversation. The company sells to health systems and billing firms through pilots, and the founders said in November 2025 that every pilot had become a paying customer.

Team of 2 to 10 · 3 months after launch

It is built for one kind of legal work, litigation, by a co-founder who practiced it. General AI chat tools have produced invented case citations that courts have punished, and Litmas AI sells a check of every citation against real court opinions as its main feature. The company stated that its demand came mainly from inbound marketing, and it priced per case so that small and contingency-fee firms can start small.

Solo founder · 7 months after launch

Designers often want a second opinion before a client review and have no colleague to ask. The maker is a product designer with more than 25 years of UX work, so the critique format follows how a senior designer reviews work: name what is unclear, then give three concrete changes. The first three critiques are free, and a subscription is needed after that.

Team of 2 to 10 · 7 months after launch

The founder, a lawyer who teaches negotiation, said that mediation works but is expensive and tied to office hours and calendars. Adri lets both sides answer at different times, and it makes the first proposal itself. The company then moved its focus from consumers to mediation providers that want a cheap first step for small disputes.