I kept hitting the same bottleneck across new projects. Coming up with a good name felt like a creative exercise, but the real constraint was more practical. A name only works if the domain is available, and checking availability mid-brainstorm breaks the flow entirely.
At first glance, this seemed like something existing AI assistants should already handle. They brainstorm well enough, but none of them close the loop on real domain availability in the same conversation. The problem is less about generating ideas and more about validating them without context-switching.
That gap made it worth building. I had already invested in reusable infrastructure in my monorepo for procedures, AI tool execution, and MCP, so this became the first fully independent product built on those shared parts.
The app runs on Supabase auth, shared oRPC procedures, AI tool calling, and a custom MCP authentication flow tied to Supabase identity. The more interesting outcome is less the product itself and more the confirmation that the shared infrastructure holds up when pointed at a real use case.