Startup Autopsy Laboratory
Your idea has a fatal flaw.
Find it for $1.
Six months from now, you'll have either a product or a lesson. Four AI examiners tell you which — before you build. Verdict in 60 seconds. No account.
Latest verdict
You're entering a proven-but-crowded $99 research-report market where LeanValidate already ships in 30 minutes—your survival depends entirely on the traction sprint tier and channel partnerships nobody's locked up yet.
Case no.
4471
Examiners
4 agents
On record
1,362
Status
Examination complete
Subject — idea submitted
“An app that reminds you to drink water”
Preliminary scan — agent findings
Market
No moat detected
Tech
Commodity execution
Finance
Negative unit econ
Timing
Market saturation
Cause of death— tap to reveal
“There are 200 of these on the App Store. The good news ends there.”
No account · No signup · Verdict in 60 seconds · $1 flat · 2 min vs. 6 months · 2-min walkthrough
Buying for a team or friend? Get a case bundle →
Live demo
What a roast actually does to an idea
Idea in. Fatal flaw out. Pivot clear.
“An app that reminds you to drink water and tracks your daily intake”
There are 200 of these on the App Store. The top three are free. Apple Health tracks this passively. You're asking someone to change a habit for a problem they've already accepted.
"Daily water tracker for kidney stone patients."
Their doctor told them to drink 3L. They have a medical reason to care. Charge accordingly.
Loops automatically · yours takes 60 seconds
Free preview — no payment needed
How good is your idea, really?
Recent verdicts
View all →“Business idea: Better Trajectory is an AI-assisted market research and validati…”
You're entering a proven-but-crowded $99 research-report market where LeanValidate already ships in …
“SaaS web app. A Qur'an memorization tool for people who memorize Qur'an but don'…”
You're building Jawhar minus spaced repetition plus a journal — that's a feature, not a company, unl…
“An app that suggests rewrites to a given resume to better match a given job desc…”
You're rebuilding Ajusta without knowing Ajusta exists, and your only differentiation — auto-apply —…
The full range
Not every idea gets roasted
The verdict depends on the idea. Here's what each outcome looks like.
“IFTA compliance software for owner-operator truckers — automates fuel receipts, mileage logs, and quarterly state filings. $49/month.”
“Mandatory compliance, captive audience, no modern SaaS incumbent. The regulatory moat is real.”
Market
IFTA compliance costs owner-operators 8–12 hrs/quarter. TruckingOffice and Q7 are legacy desktop. No modern SaaS owns this niche.
Finance
$49/mo with near-zero voluntary churn — compliance software is involuntary. CAC via trucking Facebook groups runs sub-$40.
Why it passes
Regulatory moat, near-zero voluntary churn, reachable niche via trucking communities. No fatal assumption in the analysis — the panel cleared it.
“AI meeting summaries for solo consultants — auto-generates client-ready notes from recordings. $12/month.”
“Right pain, wrong market. Otter.ai owns the generic case. The moat is in the liability, not the transcription.”
Market
Generic meeting notes is saturated. But financial advisors and lawyers need compliance-grade records — a legal audit trail, not just a summary.
Timing
SEC and FCA record-keeping requirements tightened in 2024. Regulated consultants are actively shopping. Window is open.
Pivot
Narrow to compliance-grade records for regulated advisors (financial, legal). Double the price. The audit trail is the product, not the summary.
“A social network for dog owners — share photos, find local playdates, discover dog-friendly venues.”
“BarkHappy, Dogster, and Meetup.com already lost this war. Facebook Groups won by default.”
Market
Five funded attempts at this exist. All pivoted or shut down. Facebook Groups solve the core use case for free with the existing audience already there.
Finance
Niche social networks require critical mass before any monetization. CAC exceeds $80 before the first dollar of revenue. Math does not work.
Full panel — example output
Here's exactly what $1 gets you
A real analysis, shown in full. Every agent finding. Every red flag. Every action. Nothing cut.
Subject — idea submitted
“A subscription app that teaches kids to code through interactive games — $9.99/month”
The verdict
“Code.org has $70M in backing and is free. You're selling what they give away.”
Market is real but brutally crowded. Scratch, Code.org (backed by Bezos and Zuckerberg), Tynker ($130M+ raised), and Khan Academy dominate with free products. Your $9.99/mo is competing against free.
Technical execution is straightforward — no moat in the stack. A two-person team at any of the 30+ existing edtech players could replicate your MVP in a sprint. Interactive coding games are a solved UI problem.
CAC for children's apps through paid social runs $35–80. At $9.99/mo, payback is 4–8 months. Churn in kids' apps is brutal — content requirements are constant and parents cancel when interest wanes after week three.
Edtech had its window during COVID lockdowns. The spike that launched Duolingo's explosive growth and pushed parents to pay for home learning tools has largely closed. School districts are now actively pushing back on screen time.
Competitors found during analysis
Live dataCode.org
$70M+ raised
Free. Backed by Bezos, Zuckerberg, Gates. 70M+ students. Your primary obstacle.
Tynker
$130M raised
Paid curriculum. Already in 100,000 schools. Direct competitor with a decade head start.
Scratch (MIT)
Non-profit raised
Free forever. 100M registered users. Sets the price expectation: zero.
Cause of death
Free competitors with massive backing
The market leader charges nothing. Overcoming free requires a 10x experience advantage or a different distribution channel entirely.
Post-COVID edtech churn
Retention data from the 2021–2023 edtech boom shows median churn of 40–60% after month two once novelty wears off. The TAM is real; the LTV is not.
Wrong decision-maker
Children don't pay. Parents pay — and they are already overwhelmed with subscription fatigue. The gap isn't better coding games; it's trust and habit formation.
⚠ Blind spot
You're building for the child but selling to the parent. The parent's actual fear isn't "will my kid learn to code" — it's "will they actually use this after day four." No amount of game design solves the abandonment problem without a retention loop aimed at the parent, not the kid.
Recommended intervention
Call 5 after-school program directors this week — not to sell, to ask what they spend on curriculum and what problem they can't solve. You need 3 "yes, that's real" signals before building anything.
Price one district pilot at $5k for 90 days, all-in. Remove the budget objection so you can test the actual product.
Kill the parent-facing subscription page. Every hour spent on B2C conversion is a distraction from the channel that can actually scale.
Map the procurement cycle for one specific district type (e.g. Title I schools — federal funding available, underserved, motivated). That's your wedge.
Sign up for Duolingo for Schools and spend 30 minutes understanding exactly what they don't offer. That gap is your pitch.
Or see a free sample first
Case files on record
Recent examinations
Subject
“Business idea: Better Trajectory is an AI-assisted market research and validation platform for early-stage B2B SaaS, AI, and tech-enabled service founders. The platform helps founders turn an uncertain business idea into a structured, testable market hypothesis before investing heavily in product development. Target customer: English-speaking, pre-product or pre-revenue founders who have a reasonably defined business idea but lack the time, structure, or research expertise to evaluate it properly. Problem: Founders often build based on intuition, fragmented online research, or generic AI advice. They struggle to: - Identify the best target customer - Understand how customers solve the problem today - Evaluate competitors and alternatives - Estimate market potential - Develop realistic pricing hypotheses - Find evidence of customer pain and demand - Decide what assumptions require real-world validation - Turn research into interviews, offer tests, and acquisition experiments Existing AI idea validators often produce generic reports without showing sufficient evidence or helping founders conduct real customer validation. Solution: Better Trajectory combines automated internet research with a structured, founder-led validation workflow. The system researches public sources such as competitor websites, product reviews, forums, Reddit, Hacker News, YouTube, search trends, industry publications, public statistics, and online communities. It produces: - Problem and idea assessment - Recommended customer segments - Jobs-to-be-done hypotheses - Public demand signals - Customer pain language - Competitor and alternative analysis - Competitor pricing benchmarks - Market-size scenarios with visible assumptions - Positioning and pricing hypotheses - Potential acquisition channels - Risks, contradictions, and evidence gaps - Customer interview scripts - Outreach messages - Landing-page and offer-test plans - Go, revise, or stop recommendation The platform does not claim that internet research proves demand or willingness to pay. Founders conduct customer interviews, pricing tests, outreach, and sales experiments. The system helps them prepare, analyze the results, and decide what to do next. Commercial products: 1. Idea Evidence Check A free or low-cost automated assessment that evaluates idea clarity, competitors, public demand signals, major risks, unsupported assumptions, and recommended next steps. Expected price: Free to ₹1,499 or $0–$19. 2. Guided Market Validation A sourced market research dossier combined with a founder-led validation workflow. It includes competitor research, customer pain analysis, market-sizing scenarios, ICP recommendations, pricing benchmarks, interview preparation, transcript analysis, evidence scoring, and a validation plan. Expected price: ₹7,500–₹15,000 or $99–$199. 3. Guided Traction Sprint A structured program that helps founders turn validated evidence into an offer and test potential acquisition channels. It includes positioning, offer design, lead criteria, outreach sequences, landing-page recommendations, channel experiments, funnel metrics, and weekly analysis. The founder remains responsible for outreach, sales, and execution. The product does not guarantee customer acquisition. Expected price: ₹20,000–₹40,000 or $299–$599. Differentiation: Unlike generic AI idea validators, Better Trajectory: - Provides source-linked evidence - Separates facts, assumptions, and inferences - Identifies contradictions and missing evidence - Personalizes recommendations around the founder’s constraints - Connects desk research to customer interviews and experiments - Helps analyze real customer responses - Supports progression from idea to evidence to traction - Produces an explicit go, revise, or stop recommendation Revenue model: - Free idea assessment for lead generation - Paid automated reports - Guided validation programs - Guided traction programs - Future subscriptions for founders, agencies, incubators, and accelerators Primary acquisition channels: - Founder communities - LinkedIn content - Startup newsletters - Incubators and accelerators - Development and design agency partnerships - Founder referrals - Research-led SEO - Free idea assessment tools Key assumptions to evaluate: 1. Founders consider structured market validation important enough to pay for. 2. Founders prefer a guided workflow over conducting research independently with ChatGPT. 3. Source-linked evidence and contradiction analysis create meaningful differentiation. 4. Founders will conduct interviews and experiments when given structured guidance. 5. Customers will pay ₹7,500–₹15,000 for guided market validation. 6. Validation customers will progress into the guided traction product. 7. The system can provide enough value using publicly available internet data. 8. The product can be delivered with limited human involvement. 9. The target customer can be reached economically through online”
Cause of death
“You're entering a proven-but-crowded $99 research-report market where LeanValidate already ships in 30 minutes—your survival depends entirely on the traction sprint tier and channel partnerships nobody's locked up yet.”
Subject
“SaaS web app. A Qur'an memorization tool for people who memorize Qur'an but don't speak Arabic and feel bad because they don't understand what they're memorizing. The tool makes people learn ten ayahs at a time. They See ten ayahs in arabic, and they have access to the translation of the passage, and the tafsir of each ayah from different tafsir books all in english. From there, after they've memorized, they check a box to say they've memorized the passage, and then they type what the wisdom and application they've learned is, and how they've applied it in their life if applicable, then they can move on to the next ten ayahs they want to learn. There are some progress statistics and it's meant to be a slow process.”
Cause of death
“You're building Jawhar minus spaced repetition plus a journal — that's a feature, not a company, unless the journal becomes the product.”
Subject
“An app that suggests rewrites to a given resume to better match a given job description. It then automatically applies them, so users have an apply-ready, tailored resume in less than a minute.”
Cause of death
“You're rebuilding Ajusta without knowing Ajusta exists, and your only differentiation — auto-apply — is a feature they'll ship in a quarter.”
0 cases on record
Standard examination protocol
How it works
INTAKE
01
Submit the specimen
Describe your idea — what you're building, who it's for, where you stand. Rough is fine. The agents do the research.
EXAMINATION
02
Four agents convene
Market pulls live competitor data. Tech assesses execution risk. Finance runs the unit economics. Timing checks if the window is open.
VERDICT
03
Cause of death filed
A survival tier, a one-line verdict, the strongest case for your idea, the top kill shots, the blind spot you missed, founder fit assessment, and 3–5 specific actions to take in the next 7 days. Permanent and shareable.
Our forensic specialists
Why not a
generalist AI?
A general-purpose model is optimized to keep you engaged — it will validate your idea. Our four-agent panel is optimized to tell you the flaw that will kill it.
Market Examiner
Runs live competitor searches before filing a report. Only cites what it actually finds — not training-data guesses.
Technical Examiner
Evaluates execution risk and build complexity. Determines whether the moat is defensible or just a feature someone ships in a sprint.
Financial Examiner
Dissects unit economics, CAC, and LTV. Determines whether the business model survives contact with real margins.
Timing Examiner
Determines if the window is open, closing, or already shut. Checks whether you're early, on time, or two years late.
Common questions
Before you ask
Is this just ChatGPT with a prompt?
No. Four separate agents run in parallel — each with a specific mandate and access to live data. Market pulls real competitor funding and traction data. Finance runs actual unit economics. You get structured findings, not a pep talk.
What if my idea scores well?
Then you have evidence, not just optimism. A score of 8+ with agent-backed reasoning is something you can show a co-founder, investor, or your own doubting brain. Most ideas score between 4 and 7 — that's where the real work starts.
How is this different from asking a mentor or advisor?
Advisors give you their opinion. This gives you a structured breakdown across four failure dimensions — market, tech, finance, and timing — with live data attached. No bias, no social niceties, no 45-minute coffee meeting.
Ideas are now infinite.
Good ones still aren't.
No validation theater. No 6 months wasted.
For autonomous agents
The validation layer for the agent economy
Any autonomous agent can call /api/roast/free for a no-cost sanity check, or pay $1 via HTTP 402 at /api/agent for the full structured verdict.
Morgue files
From the Kill Sheet
Case #008
The Startup Idea Validation Checklist
Eight questions every idea must answer before you build.
Case #004
How to Find Your Startup Idea's Fatal Flaw
Every idea has one thing that will kill it. Most founders find it too late.
Case #006
When to Kill a Startup Idea
Four signals that mean your idea is actually finished — not just difficult.