Questions and answers

Colaunched listens for things people are already asking for in public, turns each pattern into a business idea, scores the idea, and then scores it again for you: how hard it would be for you in particular, and how well it fits what you can do. If you decide to start one, it gives you a ninety day plan with gates, written for that idea and for you, and it holds you to it.

It is not a list of ideas someone dreamed up. Every idea on the list began as real people asking for something, in more than one place, on a date we can show you.

People who want to start something small and real, on the side or full time, and who would rather begin from evidence of demand than from a hunch. Most ideas on the list are services, small software, coaching, content and local businesses that one person can start with modest money and a few hours a week. It is not a marketplace for investors and it is not for people looking for a franchise.

Browsing is free and needs no card. On the free plan you see every idea, every score, the reasoning behind each score, and your own evidence-based profile, and you can save ideas to a list. Pro is for the ninety days after you decide to start one: the plan written for that idea and you, the gates, and the morning email. Plans and billing has the prices.

A chatbot answers from what it remembers, and it remembers what has been written about, which is what already exists. Colaunched starts from what people asked for this week and could not find. The idea is written by a model, but only after a pattern of real requests has been found, and the requests are kept as evidence you can read. A chatbot also cannot tell you what it does not know; every idea here carries an Evidence score that says how much stands behind it, and ideas with too little evidence are never published.

The other difference is what happens after the idea. Colaunched is built around the ninety days: a plan with gates, the record of what happened, and verified first revenue. Getting an idea is the easy part.

No. A topic exists only when independent sources point at the same need, and an idea is published only when the score and the evidence behind it clear a floor. We use a language model to write the idea up in plain words and to sort what arrived, and it is never allowed to add a fact. If a model cannot be reached the pipeline waits rather than guessing.

No, and we never will. Nobody can know what your business will earn, and a product that pretends to is selling you a number. What we show is Potential, which describes the size of the market we observed: how many people are asking, and what others charge for something like it. What you make is yours to find out, and verified first revenue is the first honest number.

From public demand. Every day we collect requests journalists publish when they need an expert, questions people ask in public forums that nobody has answered, one and two star reviews of the apps that lead their category, new rules published by regulators that create an obligation, what people are searching for, and what is being published. Each item is kept exactly as we found it, with the date.

When several unrelated items point at the same need, that is a topic. One request is an anecdote. Two independent ones is the beginning of a pattern, and that is the bar a topic has to clear before anything else happens.

Because the point of the evidence is the pattern, not the people in it. A journalist's request or a forum question belongs to the person who wrote it, and pointing a thousand members at one post would not be fair to them. So the idea page shows what we observed (asked for fourteen times, searches up, nobody sells a packaged answer) and how we know it, but never the names of the people, publications or vendors we observed. The same rule protects you: nothing about you is ever shown to another member.

Three gates, in order. The score and the evidence behind it must be above a floor. The signals must come from at least two independently operated sources, which is the rule that stops one mailing list from manufacturing a market on its own. And a check must be confident that an ordinary person could pursue the idea lawfully; a confident no retires it, and anything uncertain stops the machine and waits for a person to decide. Uncertainty is never resolved by rounding.

Signals are collected around the clock. Topics are formed, scored and published several times a day, and every idea is re-scored every morning. If you turn on the morning email you get the new ideas that fit you at about nine in the morning New York time, at most three a day, and never one you have already been sent.

Ideas expire. A journalist's request closes, a rule takes effect, a complaint gets fixed in the next release, a competitor appears, or the demand simply fades. When the window is closing the idea says so, and when it has closed the idea is retired rather than left on the list looking fresh. If you had saved it, it stays on your saved list and its page says the window has closed. If you had started it, your launch keeps going; the plan is yours.

That fourteen separate public items we collected, from more than one source, point at this need. A count is only ever of things we actually observed and kept. It is not an estimate and it is not extrapolated from a sample.

Sources, yes: write to hello@colaunched.ai with where the demand shows up and we will look at whether it can be collected honestly and at what it adds. Ideas of your own, not yet. An idea on the list has to be built from public demand so the evidence can be shown, and there is no way to do that for something one person believes in. You can, of course, use the evidence and the plan for any idea you like once you have started it.

Ideas are filed into broad categories of the kind a magazine section would use, such as Personal finance, Health, Small business, Parenting, Home, Technology, Food, Travel, Careers and Pets, so you can browse by the kind of buyer. When a signal arrives that the sorting step could not place with confidence, it is filed as Unsorted rather than guessed at, and a person looks at those.

The number out of a hundred is a reading of the evidence on a given day. Eighty and above we call Strong, seventy and above Good, and the rest Fair. Under it, the idea page shows exactly which parts made it: Demand, Timing, Market gap and Evidence are about the idea, and Fit and Difficulty are about you. The page also says where the idea ranks among all the ideas scored for you.

A score is not a verdict on whether the business will work. It says how much demand we saw, how open the window is, how poorly the need is served today, and how much we are sure of.

Demand is how many people are asking and how urgently, across how many independent sources. Timing is whether the window is open now: a request with a deadline, a rule that takes effect on a date, a search that is rising. Market gap is how badly the need is served: who already sells something for it, what they charge, and how the advice that exists holds up. Evidence is how sure we are of all of the above, and it is the one that stops a thin idea from being published however exciting it looks.

How well the idea matches what your profile says you bring, lack and prefer, together with the money and hours you said you would put in. It is computed for you alone and shown to no one else. Until your profile is confirmed it is blank rather than guessed, and it changes when you change your profile.

Difficulty is what it would take to get this idea to first revenue, read against what you have already done. A migration service is Low difficulty for someone who has done migrations and High for someone who has not. The note under it says which capability the idea needs that your profile does not show yet. Difficulty is about the road, not the destination; a Low idea can still have Modest potential.

Potential describes the size of the market we observed, in words: Modest, Solid, Large. It comes from how many people are asking and what others charge for something like it. It is not a forecast of what you would earn, and we never put a dollar figure on it because any figure would be invented. What a business makes depends on the person running it far more than on the idea.

Because the evidence did. Every idea is re-scored every morning against what arrived since: new requests raise Demand, a deadline passing lowers Timing, a competitor appearing lowers the gap. The idea page always shows the reasoning as of today. If you started the idea, a change in its score does not change your plan, but the launch page tells you and the next gate takes it into account.

The counting is done by ordinary code: which items point at the same need, how many sources, what changed since yesterday. A language model is used for language: to sort items into categories, to write an idea up in plain words from the signals, to check that an idea is lawful to pursue, and to write your plan. It is never given the job of inventing evidence or deciding what you will earn, and its answers are checked against rules before they are used. When a model cannot be reached the pipeline holds rather than guesses.

From evidence rather than a quiz. You can give us a LinkedIn URL, a résumé, a website, and the public usernames of accounts where people already follow you, or answer a few questions instead. From that we draft plain lines: what you bring, what you lack, what you would rather not do. Each line says where it came from, and every one is yours to edit or remove. Then you tell us your limits: the most you would put in, the hours a week you have, and how long you would give it.

It is read once to draft the lines on your profile, and it is not stored. We keep only the date we read it. The reading is done by a provider that does not keep or train on what it sees; your résumé is never sent to a free model tier that learns from its input. The lines that come out are yours to change, and they are all that remains.

Reach is how many people already follow you and what they follow you for, because an idea that needs an audience is easier for someone who has one. For the usernames you give us we look only at what is public: the follower count and the topic. We never post, never read messages, and never store your followers. Counts are observed the day you enter them and refreshed about monthly, and you can remove an account at any time.

You. Your profile, your limits, your reach, your saved ideas and your launches are private. Fit and Difficulty are computed for you alone. No other member sees your name, your numbers or your lines. Our staff see your account only when you ask us to look at something, and never sell or share it.

Edit it or delete it on the Profile page; it takes effect on the next scoring run, usually by the next morning. Lines are drafts from evidence and they are sometimes wrong about what a job title meant. If you finish a launch step that shows a capability, a line is added for it, marked as coming from that step.

Day one begins. You get a ninety day plan (six months for the one kind of idea that needs a licence before it can sell; see below) with a handful of steps and at least one gate, and within about a minute the steps are rewritten for this idea and for you: the model is given the idea's evidence, what it needs, who buys it, and your own profile lines and limits, and it hands back steps in the same skeleton. You can edit any step, add your own, and set your own gates. The plan is the method; the wording is yours to change.

A gate is a date by which something outside your control has to have happened: a stranger paid, three people said yes, a listing went live and got a reply. A step is work you do; a gate is proof the world responded. That is why a gate is never "prototype finished" or "site live": finishing is something you can do alone, and it proves nothing about buyers.

If a gate passes with neither happening, we say so, re-rank the idea for you, and the launch page tells you what the evidence suggests next. Most ideas should be killed at the first gate, cheaply, and killing one is progress.

No, because the ninety days were never the time it takes to build the thing. They are the time it takes to find out whether anyone will pay for it, and those are different for almost every business. An inventor's first gate is not "prototype done"; it is five deposits on the drawing, a signed letter of intent from a buyer, or a waitlist that converted when money was asked for. Those are things a stranger can do in ninety days about a widget that does not exist yet, and they are exactly what the plan is written to force, because the expensive mistake is fourteen months of tooling for something nobody ordered. A deposit is first revenue, and the badge counts it.

The plan knows this. When the idea is a physical product, a device, or anything manufactured, the first step is the offer of the thing before it exists (a drawing, a description, a sample, a price) and the gate counts deposits, pre-orders or letters of intent on that offer.

There is one exception to ninety days: an idea that cannot lawfully take money before a licence, certification, registration or regulatory clearance is in hand. For those the plan is written for six months, the reason is printed on the launch page, the first step is filing for it, and the first gate counts both the filing being acknowledged and the people who have said they will pay once it is lawful. You cannot choose the horizon yourself and gates do not move once set; a gate is only worth something because it was fixed before you knew how you would feel about it.

The plan written for you needs a private model, one that does not keep or train on your profile, and if that step could not run the four standard steps stay in place and the page says so. The standard plan is still a good plan. You can edit it, and you can write to us and we will look at why the drafting did not run.

Yes, but we would not recommend it. Ninety days with gates works because it is one thing. If you do run two, each has its own plan and its own gates, and the Launches page shows both.

Usually not. Most ideas here are services and small businesses where three people in three cities are not competing, and most launches stop at the first gate. The count is there so you know, not to race you. If an idea's gap closes because it is now well served, the score will say so.

The morning email already carries new ideas that fit you. Reminders tied to your plan, a note two days before a step and the morning after a missed gate, are being built and will be on by default for members with an active launch, with a switch in Settings.

It marks first verified revenue: a launch that got a stranger to pay. It is awarded automatically when the evidence meets the published requirements, it is checked the same way on every plan, and it cannot be bought. It can be withdrawn if the evidence is later reversed, for example by refunds, or if a source turns out to be false.

By evidence, in classes. Observed means we saw the payment through a read-only connection to a payment processor, a bank aggregator or accounting software. Attested means a CPA letter or tax document. Reported means you told us. Only the tier you reached is shown, only if you choose to show it, and self-reported revenue is always marked as such. We never move money and never see card numbers; we keep the totals needed to decide a tier and the class of the source, not the transactions.

Connecting a payment account is not open yet. Until it is, the Launches page records what you tell it, marked as reported, and the badge waits for an observed or attested source.

No. First revenue is checked the same way on every plan and no tier sits behind a higher price. Pro is for the ninety days; the badge is for what happened in them.

Free is for finding what to start: every idea, every score and its reasoning, your evidence-based profile, a saved list, and the morning email if you want it. Pro is for the ninety days: starting an idea, the plan written for that idea and you, the gates, and the record of what happened. Pro costs $348 a year, which is $29 a month, or $39 month to month.

From Settings, under Plan, choose Change or cancel; it opens Stripe's page for your subscription. Cancelling takes effect at the end of the period you have paid for, and your launches, plans and records stay on your account on the free plan. You just cannot start a new one until you are on Pro again.

Not for the unused part of a period, unless the law where you live requires one. If something on our side went wrong, write to hello@colaunched.ai and we will put it right.

By Stripe, on Stripe's own page. Your card never touches our servers. Changing card, changing between monthly and yearly, and cancelling all happen on Stripe's portal, reached from Settings.

Partner is for doing it with someone: matching on the gap your idea needs, fit and difficulty scored for the team, ownership and roles recorded, and verification for every member. It is on the pricing page at $59 a month billed yearly or $69 month to month, per person, and it is not open yet. Pro members will be told first.

Referrals are months, not money: share an idea, and if the person joins Pro you get a month of Pro. The sharing links that make this automatic are being built; until then write to us with the person's email after they join and we will credit the month by hand.

With at least thirty days' notice, and never for a period you have already paid for.

With Google, or with your email address and a code we send you. There is no password to make up or reset. Google or email, it is the same account as long as it is the same address.

Codes come from hello@colaunched.ai and usually arrive within a minute; check the spam or promotions folder first. Asking for several codes in a row hits a limit, so wait a minute before asking again. Enter the whole code from the most recent email, digits only. If it still fails, write to us from the address you are trying to use.

One short email a day, only on days when there are new ideas that fit you well, with at most three of them. It goes out around nine in the morning New York time, after the morning's scoring. Turn it on or off in Settings, or use the unsubscribe link at the bottom of any of them; that link works without signing in.

Write to hello@colaunched.ai from your account email with the subject "My data" and we send everything we hold about you, or remove it. Deletion removes your profile, reach, saved ideas and launches within thirty days; server logs age out on their own within ninety. The buttons for both are coming to Settings.

We use one cookie to keep you signed in and a little browser storage for conveniences like a remembered filter. There are no advertising cookies, no cross-site tracking, and no third-party tracking scripts. We do not sell data and we do not share it with advertisers. The privacy policy is written to be read.

Colaunched is for adults. You must be at least eighteen to have an account.

No. Colaunched is information and tooling: what people are asking for, how sure we are, and a method for finding out whether they will pay you. It is not investment, legal, tax or financial advice, and starting a business is your decision at your own risk. Have a lawyer look at anything you sign, especially with a partner.

Yes. The demand is real, because we observed it, but the write-up is a reading of it and readings can miss. That is what the gates are for: the first one is designed to find out cheaply whether the world agrees. If you think an idea is wrong, or unlawful, or unfair to someone, tell us; a person looks at every report.

You can use any idea for your own launch and, once sharing is live, share individual ideas through the app. Copying, scraping or reselling the list or its evidence is not allowed, and neither is using automated tools on the service other than the ones we provide.

Cyligence, LLC, a small company. The fastest way to reach a person is hello@colaunched.ai, and a person answers.