Solo developer in Vietnam. I pick a narrow, boring problem that costs somebody real money, ship something small, and talk to the people it is for. Some of it works. I write down the parts that do not.
Advance warning of third-party API breaking changes for automation agencies. Watches the official changelogs and deprecation pages of 16 vendors, maps announced changes to the n8n / Make / Zapier workflows an agency runs for each client, and sends a per-client digest before the deadline hits.
Piracy monitoring for online course creators. Finds leaked copies of paid courses, tracks where they spread, and tells the creator what is actually at risk. Sold worldwide through a merchant of record.
Tracks pricing and plan changes across SaaS vendors and publishes them as searchable pages, so anyone comparing tools can see what a vendor charged before, not just today.
Narrow beats big. Every tool above serves one kind of person with one recurring problem. I would rather be indispensable to 60 people than mildly interesting to 6,000.
Sell before building. The payment page goes live before the machine does. If nobody pays in the first weeks, I write up why and move on, instead of building for another year in the dark.
Say what is actually true. Coverage pages, honest limitations, real dates. Monitoring is best-effort and I say so on the pricing page.