Startup ideas your AI agent can build
Every idea ships with a UIQ score you can audit, the reasoning behind it, and an agent-ready build kit that ends with a running MVP.
Free forever · Idea of the week fully unlocked, no account needed
Launch offer40% off annualPro at $89.40, was $14913 days leftExpecta · Product safety, scanned
Idea of the week
- Demand80
- Feasible78
- Codeable76
- Timing76
Locks Monday, then the free sections stay.
From first prompt to a running MVP
Every launch-ready idea ships with a ten-phase prompt sequence you run on your own AI agent. Brief it, lock the spec, scaffold, build, verify, deploy. The checklist tracks you the whole way.
- 8 sequenced prompts, each with a verification step
- Works with Claude Code, Cursor, Lovable, or Bolt
- 2-3 stack options with per-stack difficulty ratings
- Progress saved across sessions, weekly nudge email
Expecta · Build checklist
3/10 · 30%An operating system for the whole journey
AI made writing code the easy part. What kills projects comes after: the wrong idea, no customers, quitting too early. Your dashboard is built against all three.
Validated idea
Start from evidence, not a guess. Every idea arrives scored, researched, and ready for your agent.
Running MVP
The kit walks your agent from brief to boot, one verified phase at a time, blockers pre-solved.
First users
The launch kit maps your exact channels and scripts your first ten user conversations.
$1K MRR and beyond
Momentum, traction, and finance panels keep your pace honest against the market path.
Latest idea reports
Seedbox · Test data that looks real
Developers need realistic data to build against. The options are a production dump, which is a breach waiting to happen and increasingly illegal, or generated nonsense that hides every bug caused by real-world shapes. So most teams quietly use the dump.
Expecta · Product safety, scanned
New parents make dozens of small safety calls a week, is this lotion fine, is this snack an allergen risk, and answer them by pasting an ingredient list into a search box and reading ten conflicting forum threads. The information exists but it is scattered, unranked, and never tuned to their own child.
Marklight · Prove what your AI made
An agency ships thousands of AI-assisted images, clips, and voiceovers a month for clients who are about to be legally required to declare exactly that. Nobody can currently answer the simple question of which asset was machine-made, by what, and when, because the answer was never written down at the one moment it was true.
From signal to shipped product
Evidence, not opinions
Every idea starts from real complaints and search demand, with cited sources you can open and judge yourself.
Scores you can compare
Eight sub-scores on the UIQ rubric, one comparable 0-100 verdict across the whole database.
Execute the same day
Run the ten-phase kit with your agent and check off the steps until the MVP runs live.
Inside every idea page
Identity
Project name, essence, category, pitch
UIQ score
0-100 verdict with 8 sub-scores
Problem and evidence
Cited pain: who hurts and how badly
Solution and scope
The product, MVP scope, non-goals
Target user
Persona, where they gather, who pays
Founder fit
Difficulty, solo score, time to MVP
Market and timing
TAM, SAM, SOM and the why-now
Competition
Real players, pricing, and the gap
Business model
Pricing, unit economics, $1K MRR path
Business plan
Channels, cost structure, 90-day GTM
Financial projections
24 months, 3 scenarios, break-even
Investment potential
Bootstrap or fundable, comparables
Risks
Top 5 scored, each with a mitigation
Strategies
Growth, pricing, first 10 users
Vibe-coding kit
10 phases from first prompt to MVP
Extras
Domains, landing copy, changelog
Browse by category
What the scores actually say
What early users say
“I picked one idea, ran the kit phase by phase, and had a working MVP in nine days. The checklist alone saved me a month.”
“The kit prompts are better than mine. The verification steps caught two things I would have shipped broken.”
“The UIQ scores made it easy to kill three weak ideas fast and commit to the right one. That alone was worth the year.”
Start free, upgrade when you commit
Free
- Idea of the week, fully unlocked
- UIQ score, problem, and founder fit on every idea
- Weekly idea email
- 1 active build · save 3 ideas
Pro
- Every validated idea, unlocked in full
- Kits + AGENTS.md tuned for Claude Code
- How it makes money, the risk, the build notes
- 3 active builds at once · the idea finder
Max
- Everything in Pro, plus:
- Early access: new ideas 7 days before Pro
- Priority kit-fix SLA
- Unlimited builds · the retired-ideas vault
Frequently asked
How are ideas researched?
Each report starts from live web evidence, complaints, reviews, and search demand, with every claim cited so you can weigh the reasoning behind each score.
What do I get for free?
The idea, UIQ score, problem, target user, and founder fit on every idea, plus the full idea of the week unlocked for the whole week. The rest is Pro.
Which AI agents do the kits work with?
Any capable coding agent: Claude Code, Cursor, Lovable, or Bolt. Every phase is a copy-paste prompt with its own verification step before you move on.
Do idea pages go stale?
No. Pages are living documents: competitors, trends, and UIQ scores get refreshed over time, and every update lands with a visible changelog entry.
What does the score actually mean?
UIQ is one comparable 0 to 100 verdict from eight sub-scores, so demand, competition, and buildability are weighed the same on every idea. GO is 75+.
Do I have to pay to start building?
No. Browse the whole database, read every idea core, and start a build on the free plan. Upgrade only when you want the full kit for your chosen idea.
Your next product is already validated
One validated, agent-ready startup idea in your inbox every week.
Free forever · idea of the week fully unlocked · no card