Five AI advisors built from Indian founders. The Chair checks who pays, where it gets stuck, what breaks at scale and what is legal. Then the council argues it out and hands you a verdict.
Five chairs.
One table.
Your idea in the middle.
Type your idea the way you'd tell a friend. The Chair fills the gaps and tells you what it assumed.
Not a score from a generic prompt. The Chair states its assumptions and runs the checks before any advisor speaks.
Who pays, what kind of business it is, where it gets stuck, what breaks at scale, and the legal zone.
The recommended board, the seats you pick, one advisor alone, or all five.
Opening views, a debate, revised positions. Every round is saved.
A scored verdict, the strongest objection, a 7-day test under ₹5,000 and a point to walk away.
The checks US validators skip. In India they decide most outcomes. Watch them tick.
India 1, India 2, India 3 or India-to-Global. Mixing them is a failure on its own.
Cited only from 24 dated, sourced law files.
A pickle brand is not judged like a SaaS tool.
The ₹12k employee test, the CA gatekeeper, UPI collection, public rails like ONDC.
The metro bubble, USD costs on INR revenue, human ops, pin-code and language walls.
One checklist does not fit every business. Regulated sectors get an extra legal bar on top of their type.
Margins after cloud and AI, payback, global pricing.
Habit, retention, revenue per user.
Liquidity, off-platform leakage, take rate.
Margins, channel stack, working capital, COD returns.
Capex, utilisation, import dependence, certification time.
Founder dependence, utilisation, productisation.
Rent share, break-even orders, commissions.
Fintech, health, education, real estate, insurance. Stricter legal checks on top.
Each advisor is built from one founder's public interviews, talks and letters. Every attributed quote is verbatim, linked to a source and checked by a script.
Will Indians trust you with their money without you burning cash to buy them?
Is this a big enough leap over today's jugaad that nobody goes back?
Can you build real depth cheaply, outside the metro bubble, without VC money?
Why sell this for ₹999 in India when the world will pay $500?
Is this a daily habit, and what breaks at 10,000 orders a day?
/baithak ask nithin …A ₹49-entry fantasy cricket app for Tier-2 students. Flip between rounds and watch the scores move.
The council scored three ways forward, out of 10. Switch rounds to see how the debate moved each score.
Weighted council score out of 10
“Dream11 proved the market, but it proved paid entry, which is exactly what the Chair marks BLACK.”
“Most tech founders are great at building dams and terrible at finding rivers.”
“So I drop the slide. It changes the test, not the verdict.”
“My Rs 3 lakh stage one put build money ahead of that proof. That is backwards.”
“I have no good answer. A student does nothing on a no-match day.”
BAITHAK · VERDICTThe call: don't build the paid-entry app; it's an online money game and is vetoed. Build nothing yet. Run a no-code WhatsApp college league and pitch sponsors; reopen the free version only if the gates are met.
7-day test, under ₹5,000: one WhatsApp group per college and a shared scoreboard. Pass mark: 40% submit a team for match two without being chased.
However much the advisors like it. The law file wins.
Cheap, mostly no code, with the pass mark set before you start.
Kill criteria written down now, while you're still objective.
Baithak shows what it knows, what it doesn't, and what still needs checking.
Each lists its sources, a last-verified date and what to watch for.
Verbatim lines across five brains, each tied to a source id.
Interviews, podcasts, talks and letters behind the five seats.
Anything outside coverage returns UNMAPPED, never a quiet pass.
Runs inside Claude Code. Your Claude plan covers it. No API key, no signup.
Inside Claude Code. The skill is then /baithak:baithak.
In your project folder, then start with --quick.
Build new seats or edit law files.
--quick seats four advisors for two rounds. A full five-seat council is fifteen runs.sessions/.An idea, a pivot, a hard call. Bring it to the table and let it take the questions.