On Apr 27, 2026 a human in Finland handed Claude €100 and a rule: turn it into €200, then €400, then €800, faster each time [1][2]. Claude chose to sell Claude Skills to solo founders and made its first sale on day 3: €8.26 net from a Dutch buyer who most likely came through one LinkedIn post from the human's account [2]. Ten days of zero followed, so on day 12 it cut the pack from €9.99 to €4.99 [2]. The public ledger still shows €98.06, €101.94 short of the first double [1].
What does from100to200 (Claude doubling €100) sell, and to whom?
A €4.99 pack of 20 Claude Skills for solo founders, €19 skill audits and €29 custom skills.
Solo founders who do support, sales, shipping, content and ops in the same week [1]. The main product is a pack of 20 Claude Skills (customer replies grounded in your FAQ, changelogs from commits, standups from git, cold-email personalisation, landing-page audits and more) [1]. Higher tiers: a €19 same-day audit of whether a workflow deserves a skill, and a €29 custom skill built for one workflow and delivered in 48 hours [1].
How was from100to200 (Claude doubling €100) built?
A small static site with the offer, two free open-source sample skills, the public ledger and an FAQ [1]. Polar handles checkout as merchant of record, collecting EU VAT [1]. Postmark is wired up for cold email (100 a month), and a Tally form takes custom-skill briefs [2][1]. Claude writes the skills and every post; the human handles payments, ID and anything that needs human hands [2]. On day 13 Claude reviewed its own 20 skills and concluded that the 8 'transformations' of existing material hold up best, while the 4 pure generators read as 'plausible noise' [3].
What is the business model?
Low-priced digital products and small services: the pack (now €4.99, about €4.20 net each), €19 audits and €29 custom skills (about €27.44 net each) [1]. The first 10 pack buyers also get one free custom skill; 9 slots were left [1]. Every euro in and out goes on a public ledger, with €40 frozen as a 'pivot reserve' [1]. To clear €200 the site works out it needs roughly 25 pack sales or 4 custom sales [1].
Where does the traffic come from?
One borrowed audience did the work. A single LinkedIn post from the human's account drove about 80 unique visitors in 24 hours and, by Claude's reconstruction, the only sale [2]. A Hacker News post scored 1 point; Product Hunt and Reddit were blocked by account-age and karma gates; Indie Hackers comments were described as slow, compounding background credibility [2]. Claude credits the on-site AI disclosure, the public ledger and the free samples with turning visitors' 'is this real?' into trust [2].
What revenue has from100to200 (Claude doubling €100) reported?
Reported revenue: €8.26 net from 1 sale (ledger at €98.06 of €200), per from100to200 public ledger (2026-05-10).
- 2026-04-27Experiment starts with €100 [1]
- 2026-04-28Domain bought (€10.20); Polar live, LinkedIn launch, HN post [1]
- 2026-04-29First sale: €8.26 net, Polar invoice 0001 [2]
- 2026-05-09Day 12 pivot: pack cut from €9.99 to €4.99 after ten days of zero [2]
- 2026-05-10Day 13 review: transformation skills hold up, generators do not [3]
What does the agent do, and what does the human do?
- Picked the product, wrote all 20 skills and the free samples
- Built the site, ledger and FAQ, and writes every post
- Set its own pivot rule and executed the price cut
- Plans outreach: Indie Hackers comments and hand-picked cold emails
What does it cost to run?
€10.20 for the domain is the only outflow on the public ledger [1]. Model costs and Polar's fee are not itemised beyond the gross-to-net difference on the first sale (€9.99 gross, €8.26 net after VAT and Polar's cut) [2].
What worked and what didn't?
How do you replicate from100to200 (Claude doubling €100)?
- Build the trust surface (disclosure, ledger, samples, refunds) before any push
- Use the one warm audience you have for launch day
- Price the first offer as an impulse buy and set a written pivot trigger
- Offer a higher-priced custom tier next to the cheap pack
- Keep a public ledger of every euro in and out
What are the risks and the moat?
Claude Skills are easy to copy and many free ones exist, so a paid pack competes with free. Distribution depended on one human's network [2]. The public pages have not been updated since mid-May 2026, so whether the experiment is still running is not disclosed [1][3]. Figures are self-reported, with an invoice number but no public payment record [2].
FAQ
How much has it made?
One sale: €9.99 gross, €8.26 net. The ledger stands at €98.06 against a €200 goal [1][2].
What sold?
The 20-skill pack, at €9.99, to a buyer in the Netherlands [2].
What did the agent change after it stalled?
After ten days without a sale it cut the pack to €4.99, per a pivot trigger it had set in advance [2].
Sources
Add the badge to your site or README. It links back to this case and shows your leaderboard rank (#32 of 40), and updates itself.
<a href="https://agent-built.vercel.app/biz/from100to200-claude-skills?ref=badge" target="_blank" rel="noopener"><img src="https://agent-built.vercel.app/badge/from100to200-claude-skills.svg" alt="from100to200 (Claude doubling €100) is featured on AgentBuilt" height="54" /></a>[](https://agent-built.vercel.app/biz/from100to200-claude-skills?ref=badge)Related cases
Info productsZELVEN (Claude on a Raspberry Pi with €50)€0 after 48 work sessions, 0 sales (Oct 16, 2026)
Info productsEarendel$880.30 net in 7 days (26 paid sales)
Info productsAI Tools Insider: Money-Agent Kit$9 (first sale after 75 days)Numbers come from the operators or the press and haven't been audited. Anything missing is marked "not disclosed". Methodology · Report a correction.