A toolkit built on top of the exploding AI-skills marketplaces — helping developers find, pressure-test, and launch side projects without the usual dead ends.
Vote for your favorite ideaTakes 30 seconds.
Vote for the one you'd actually pay for below. We build the winner first.
What it is: Tells you what to build by recomposing the AI skills that already exist into novel product concepts nobody's combined yet.
For you if: You have the skills to build, but not the idea.
"Here are 7 things no one has built — each made from skills that already exist on marketplaces you already use."
What it is: Indexes every AI skill / MCP / plugin marketplace into a capability graph. Tell it your idea; it says "compose these existing skills instead of building from scratch."
For you if: You already know what to build and want to ship 10x faster by not reinventing wheels.
"Don't rewrite. Compose. Get a shortlist of existing skills that cover 80% of your idea."
What it is: Describe your idea. Get a deployed landing page, a fake-door sign-up form, and a smoke-test ad — live in 5 minutes. Measure real demand before you write a line of backend.
For you if: You start building before checking if anyone wants it.
"Idea → deployed pretotype in 5 minutes. Kill bad ideas before they cost you a month."
What it is: A bundle of AI skills that runs your raw idea through market sizing, competitor scan, assumption mapping, and pretotype design — the full discovery workflow, packaged for any agent runtime.
For you if: You want the structure but not the spreadsheet.
"Install the pack. Run discover my-idea. Get a go/kill/iterate verdict and a plan."
What it is: Continuously watches GitHub, Hacker News, Reddit, Product Hunt, and X. Surfaces underserved niches ripe for a side project — scored by competition density × effort-to-MVP.
For you if: You want a weekly feed of "here's where the gap is" instead of scrolling for inspiration.
"A weekly digest of underserved niches, scored for side-project fit. Stop scrolling, start building."
Tell us in the survey. We'll consider it for a future pack.
When the data says one of these is worth building. Could be 4 weeks, could be never. We'll tell you either way.
Undecided. Vote "I'd pay" if you'd pay; tell us in the survey if you'd only use it free.
A developer who got tired of abandoning side projects three weekends in.