Candle AI: Product, team and how it was built

Team of 2 to 10 · sold 17 months after launch

Four founders, two of whom had built the drinks-delivery start-up Winelivery, looked for a profession. They wanted one that works with a large, fast-changing body of rules and had not yet adopted AI. They chose Italian accountants and labour consultants. They tested a first version with practitioners and sold monthly subscriptions. In October 2025 TeamSystem, an Italian maker of software for accountants, bought the majority of the company. Its press release said Normo.ai had €1.3 million of revenue in 2025.

Team of 2 to 10 · 15 months after launch, revenue growing

It sells one narrow job to one profession in one country: missed calls at small Belgian law firms. The site, the blog and the terms are written for lawyers at the Belgian bars, in French. They answer questions about professional secrecy and GDPR that this buyer asks first. A small firm pays a fixed 90 euros a month instead of a part-time secretary.

Team of 2 to 10 · 6 months after launch

The founders are lawyers who built the legal departments of two Berlin startups. The CEO writes that he knows the frustration with old legal software from that work. LawX replaces the whole office system instead of adding one more AI tool. It charges one price per user for the whole team, assistants included. Every AI suggestion waits for a person to approve it. tech.eu reported in May 2026 that the next step was the broader law firm market.

Team of 2 to 10 · 21 months after launch

One founder had worked as a corporate lawyer and the team sold directly to law firms and banks that it already knew. The product was built on Indian court material, which global tools cover less well. Customers made it part of daily work. When the platform went down for maintenance, users called and messaged the team to ask when it would return.

Team of 2 to 10 · sold 28 months after launch

Three founders built it without outside money in 2023, and the first message, that it replaced lawyers, scared the customers away. They changed it into an assistant for lawyers. They connected it to the Spanish official gazette and court decisions. They had it take the Spanish judiciary entrance exam in front of a notary. According to Merca2, that public test turned a handful of customers into hundreds a month. Doctrine, a French legal AI company, bought it in February 2026.

Team of 2 to 10 · 6 months after launch

The founder is a New York trial attorney with about fifteen years of practice. The company's pages say he built the platform inside his active trial practice and wrote the prompts and templates himself. General legal AI tools are not built for one task such as contested litigation; this one is built only for that. Its pricing page says usage is designed to be passed to clients as a disbursement, like research time on Westlaw or Lexis.

Team of 2 to 10 · 27 months after launch, revenue falling

Many experts know that regular LinkedIn posts bring clients, but they do not have the time or the writing habit. UpLinked sells one narrow job, about one hour a week of LinkedIn posting for business-to-business experts, first in Dutch and later in English. Coaches and consultants who teach LinkedIn used it themselves and for their own clients. The sellers wrote that this word of mouth brought every customer, with no paid ads. TrustMRR recorded US$200,908 of Stripe payments to the founding company up to October 2026. Monthly revenue fell by more than half in 2026. The business was listed for sale, and since late October 2026 another company is named as its operator.

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

The founder had already grown one tool for dropshippers, Minea, to US$750,000 a month of recurring revenue by his own account. He did it by putting a YouTube creator in the founding team. He repeated the same method. A creator co-founder makes videos that bring the customers. The product shows its result in a few minutes, and a free version charges only at the moment of publishing.

Team of 2 to 10 · 11 months after launch, revenue growing

Three of the four co-founders had LinkedIn audiences, together more than 480,000 followers, made of the people Kleo sells to. The free extension had already reached tens of thousands of users. When LinkedIn stopped the extension, they sold a rebuilt paid version to that audience first, in limited groups of seats, before any public launch.

Team of 2 to 10 · 3 months after launch

Two founders who had run a car rental side business saw that fleet companies kept their planning, tracking and invoicing in separate tools. Fleet companies lost money in the gaps between them. They sold one control center that joins operational and financial data. In May 2025 they wrote that revenue had reached US$200,000 a year, all from word of mouth.

Started small, grew after funding · sold 11 months after launch

Its founders wrote some of the methods that the industry already used to run models cheaply, such as AWQ for 4-bit weights. They turned that research into a paid serving API. They proved the speed on a public benchmark and partnered with a cloud company that needed exactly that speed. Eleven months after its first public post, the partner agreed to buy it.

Solo founder · 33 months after launch, revenue falling

The app costs almost nothing to run: the owner reports a profit margin of 99%. The App Store page shows no new version since July 2025. A small, steady flow of downloads, about 2,500 in 28 days in October 2026, keeps about 150 paid subscriptions alive. The owner is an experienced app marketer who already sells more than a dozen other apps from the same developer account.

Solo founder · 4 months after launch, revenue falling

The founder turned the go-to-market frameworks she uses as an advisor into a product with one job and one price. It is a small, early business. Stripe data shows US$500 of revenue in its first four months and two active subscriptions in November 2026. The founder has listed it for sale.

Team of 2 to 10 · 11 months after launch, revenue flat

The site ties a cruise chat product to a creator with a social audience. It says the tool is led by her experience of living, working and travelling at sea. The first verified payment came in January 2026, the month after the start date on TrustMRR.

Team of 2 to 10 · 45 months after launch, revenue falling

Said US$19K a monthRecords US$9.6K in JulyDid not hold

It appeared in January 2023 inside Google Slides, a program where people already work. At that time writing slides from one sentence was new, and users showed it to each other without paid advertising. A blog about presentation topics then brought about 80% of new users from Google search, by the founder's account. Both channels weakened: search traffic fell, and Google added the same function to Slides. The verified total is real, and so is the fall: monthly revenue fell by half between February and July 2026.

Solo founder · 7 months after launch, revenue growing

It went where the users already were. It became an app in ChatGPT's app store and a plugin in Claude's directory and in Claude Code, with setups for Cursor and Codex. It did this in the months when these directories were new. It charges per tool call instead of a share of ad spend, so a small advertiser starts at US$49 a month. And it applied to the platforms' own API programs. The company says Google approved it in June 2026 and Amazon admitted it to its partner network.

Team of 2 to 10 · 34 months after launch, revenue falling

It entered a market where people already paid for tools such as Taplio, and did the one core task better: writing the post. The founders' public posts brought the first customers and a lifetime-deal platform, and that deal paid for growth without investors. Subscriptions, partnerships with LinkedIn influencers, affiliates and a move to team plans came after.