What the agent did
Mert used Claude Code to build claudemarketplaces.com, a curated directory of Claude Code plugins, skills and MCP servers, in about two hours as a test of the feature-dev plugin. The site later sold clearly labelled ad slots.
Result, as claimed
He says he built claudemarketplaces.com in about 2 hours to test the Claude Code feature-dev plugin, and released it about 5 hours after plugins launched. One year later he reports 26M+ Google impressions, 800k+ clicks, 2.5M unique visitors, a peak of 550k+ monthly unique visitors, and $67k+ in ad revenue. A screenshot he posted shows about 807k total clicks and 26.9M total impressions.
This is the result as claimed in the public source linked below. It has not been checked and is not a typical outcome.
Proof
Try it yourself
Scope a tiny niche directory you could ship in one coding-agent session, then decide what you would sell once search traffic arrives.
Prompt
I want to ship a small public directory with a coding agent. Niche: [describe the niche and who searches for it]. List the fields each listing needs. Write a one-page build plan for a first version that can go live in one session: pages, data source, how new listings get in, and a simple ad or sponsor slot. Then list three ways this could fail to make money, and what proof I should collect in the first 30 days (traffic, clicks, enquiries).
Tools
- Claude Code Named by Mert as the tool used to build the directory in about 2 hours
Steps
- Pick a niche you already understand and can keep clean by hand at first.
- Run the prompt and cut the plan to one session of work.
- Ship a public page, then measure search clicks before you sell ads.
- Only open sponsor slots once you can show a real traffic screenshot.
Watch out: Traffic is not revenue. An ad slot only works after strangers already visit. Keep editorial control so the directory does not fill with spam.
Starter version by Solopreneur Code