Case 1: AI Resume Optimizer
A former Big-Tech HR in Hangzhou turned solo developer and did exactly one thing: rewriting resumes line by line against job descriptions using an LLM. Priced at 29 CNY per month, acquiring users through Xiaohongshu and Zhihu content, it reached 3,500 paying users in 18 months — about 1.2M CNY in ARR. All code written by the founder plus Cursor; server costs under 800 CNY per month.
Case 2: E-commerce Image Tool
A Shenzhen product manager who used to sell on Amazon productized his own biggest pain — white-background and lifestyle images. Upload a product photo, get platform-compliant image sets. At 299 CNY per year it serves over 3,000 cross-border shops, generating about 900k-1M CNY annually, with support handled by a Coze bot.
Case 3: Podcast Transcription and Summaries
A Beijing content professional productized the most time-consuming step of podcasting: upload audio, get transcripts, chapter summaries, and show notes. At 49 CNY per month plus pay-per-use, monthly revenue stabilized at 80-100k CNY. All acquisition is SEO — the product ranks first for "podcast to text" keywords.
What They Share
- Razor-narrow scope: each solves exactly one recurring pain;
- Subscriptions only: predictable revenue with 60%+ renewal;
- APIs all the way down: no self-trained models, cost risk stays upstream;
- Content and SEO for growth: zero sales team, the founder is the channel.
The Repeatable Path
| Stage | Action | Duration |
|---|---|---|
| Pick | Find a pain that recurs 3+ times a week in an industry you know | 2-4 weeks |
| MVP | Assemble APIs into a minimal product; charge first, polish later | 2 weeks |
| Growth | Publish 20 SEO pieces around the pain keywords | Ongoing |
| Scale | Push renewal to 60% before adding features | 6+ months |
Bottom line: the core of micro-SaaS is not technology — it is finding a small enough pain in an industry you know deeply. Charge first, perfect later. That was the shared starting point of all three cases.