Log every startup idea you have. SiftId gives each one a calibrated 1–9 viability score and a brutally honest two-sentence rationale — built from the same frameworks real VCs use. Most ideas score 5. That’s the point.
The pain is real and the audience is identifiable, but TAM is hard-capped at ~10% of the population and subscription fatigue makes recurring revenue brutal in this category. Pivot toward a one-time gift store before committing to recurring billing.
Every other “AI startup evaluator” is built to make you feel good. They give 7s and 8s to anything coherent. They use words like “promising” and “could be interesting if”. They optimize for the dopamine, not the truth.
SiftId is calibrated. The median idea is a 5. Fewer than 10% of ideas score 8 or higher. The maximum is 9 — because there is always uncertainty. If you want encouragement, you can find it elsewhere. If you want signal, log an idea.
Your dashboard, ranked by viability. The good ones float. The 5s give you something to refine. The 2s tell you to move on.
Real, painful, frequent problem for accelerator partners drowning in 500+ applications a cycle, with a clear path to enterprise pricing. Differentiation is thin against generic LLM workflows, so the moat has to come from the rubric and calibration data over time.
Tar-pit category — 99designs has owned this for 15 years and the unit economics for new entrants are brutal because designers commoditize each other in race-to-the-bottom auctions. Pivot toward a curated specialist network or vertical-specific design retainers.
Real founder pain (you forget the good ones, the bad ones eat your bandwidth) and a clean monetization story at $5/mo, but the differentiation has to come from calibration credibility and the freemium model has to convert. Schlep work in the rubric is the moat.
Every idea is evaluated against a weighted rubric synthesized from YC, Paul Graham, Thiel, Sequoia, Porter, Lean Canvas, and RICE. No grade inflation. No participation trophies.
Title and a sentence is all you need. Log it before it fades. Edit it later when the shape gets clearer.
Keyword extraction tags every idea by category so the patterns in your thinking become visible over time.