One-Person Company cohort
Eight weeks to build a small real product with AI, launch it, find real users and write down honestly how it went. Shipping, not simulating.
Who it's for
- University students (undergraduate or master's) in mainland China. This is cohort one; other regions come later.
- Any major. You don't need to know how to code, but you need to be willing to learn by doing.
- About 10 hours a week, for 8 weeks. If you can't commit that, please wait for a later cohort; it's fairer to the other applicants.
What you leave with
- A live product at a public link.
- Real usage data, even if the numbers are small.
- Eight honest build logs and a case study.
- Everything you built is 100% yours. STARC takes no equity, no revenue share and no credit.
No certificate. The link to your product, data and logs is the proof.
One concrete deliverable every week
Each deliverable goes into that week's build log.
| Week | Focus | Deliverable |
|---|---|---|
| 0 | Set up | AI coding assistant, git and a deploy platform working; one line of code live on the internet; 3 candidate problems |
| 1 | Find the problem | Talk to at least 10 potential users; pick one problem; one-page problem statement |
| 2 | Smallest version | v0 live at a public URL |
| 3 | First users | At least 5 real users have tried it; notes on how they used it and where they got stuck |
| 4 | Iterate + midpoint | Fix the biggest problem; 5-minute demo at group review |
| 5 | Distribution | One channel experiment (Xiaohongshu, Zhihu, a community…) with numbers |
| 6 | Retention or revenue | Measure whether users come back, or try charging, even ¥1 |
| 7 | Polish | Reliability, privacy, onboarding; write the case study |
| 8 | Demo Day | Public showcase: live product, real numbers, full build log |
Mostly async, by design
Sunday: build log due
What shipped, numbers, what users said, where you're stuck, next step, and how you used AI.
Monday: digest
An AI summary of all 12 logs, so the reviewer can focus on judgment.
Wednesday: 1-hour group review
3–4 projects in depth each week. Everyone else gets written feedback. AI-drafted feedback is labeled as such, and a person reads it before it's sent.
Anytime: peer pairs
You and a partner test each other's products and read each other's logs.
Two outcomes, both public
Shipped
Product live at Demo Day. At least 10 real users who aren't classmates, friends or family, backed by screenshots, analytics or an anonymized list. All 8 logs. A public case study.
Completed
Full participation and all 8 logs, but fewer users. Your case study explains why it didn't work and what you learned. That's valuable too.
Either way, your deposit comes back in full.
It depends on effort, not results
The deposit exists to stop people from disappearing halfway, not to punish failure. If your product finds no users but you took part honestly, you get every yuan back.
Kept only if you miss two weekly logs in a row without telling us, quit after week 1, or fake data or someone else's work. That last one also ends your participation.
Always returned if you leave during weeks 0–1, or because of illness or a family emergency. A short note is enough; we won't ask for proof.
“Nobody used it” beats a made-up number
Faked data means removal. Logs need evidence for their numbers, and peer pairs test each other's products. We'd much rather read an honest log about a flop than a polished one about nothing.
Your logs are private to the cohort by default. Each one is published only if you agree, with your name or anonymously, and you can withdraw it later.
Show us you can build, and that you'll tell the truth
The basics
Who you are and how much time you really have.
Something you've made
Code, writing, a video, a club, a small business, anything. If there's nothing yet, one page on a problem you want to solve and how you'd test it.
The 48-hour task
Use any AI tool to build a one-page tool that anyone can open, plus a note of up to 300 words on what went wrong and how you fixed it. Rough is fine; we want to see that you can ship and record it honestly.
Reviewed with school names removed. We look at three things: did you ship, is your account honest, and is it a real problem for real people.
Submit the application form now to register. We'll contact you when review begins.
We never charge before admission.