How the system works, where ideas come from, how they’re scored — and how your feedback makes blink ideas smarter over time.
blink ideas is an AI that does the hardest part of starting a business for you: finding an idea people actually care about. It continuously watches the web for real people asking for solutions that don’t exist yet, screens those conversations for genuine demand, and turns them into ranked business opportunities with a score, a target customer, competitors, and the first steps to validate it.
No account and no payment — blink ideas is free to use. You can browse ideas, vote, leave feedback, and ask the AI questions right away.
The worker runs continuously, in five stages:
The status bar at the top of the page shows which stage it’s in right now.
It shows what the worker is doing at this moment: collecting (gathering fresh conversations), processing (screening and analyzing), or idle/completed (finished a full pass). A red dot means something went wrong — it retries automatically. New ideas appear continuously, so it’s worth checking back.
Continuously. Sources are polled on a short cycle around the clock, and every pass can surface new opportunities — so the list keeps growing, and each generation of ideas is informed by everything learned before it.
Ideas are re-scored over time, especially as new community feedback comes in — a score (or the list order) can shift as blink ideas learns. Ideas themselves are never removed for being “bad”; the ranking just reflects the latest knowledge.
Currently seven communities, plus general web search:
Different communities surface different kinds of problems — a local service idea rarely shows up in Hacker News, and a developer tool rarely shows up in other niche communities. Watching many sources is what keeps the idea mix diverse.
Yes — that’s part of the long-term roadmap. As more people use blink ideas and share feedback, it keeps expanding the sources it scans and getting smarter and smarter. More sources means more problems found, and more problems found means better ideas.
No. Every idea is anchored in real demand signals — an actual person or community asking for something that doesn’t exist yet. The AI screens those conversations and only turns genuine problems into ideas. When the tools can’t find anything relevant, it says so rather than inventing something.
Yes — click any idea to open its details and you’ll see the evidence: real quotes from the conversations that showed demand, with their source and a link back. You can always read the original context yourself before committing to an idea.
Start with the three numbers: the score (how good the business looks), the confidence score (how well proven the demand is), and the evidence (read the original conversations). Then weigh what fits you: your skills, your interests, and how much capital you have. A 70+ score in a space you understand beats a 90 in a space you don’t.
That’s usually a good sign — competitors listed next to an idea are validation that people pay for this. Existing companies prove the market; your job is to differentiate. If the space feels too crowded, the competitors list makes that easy to see, and you can move to the next idea.
No — the business format always matches the problem. That means local services, physical products, marketplaces, agencies, subscriptions, and coaching businesses all show up, not just software. Software is only chosen when it’s genuinely the best way to solve the problem.
No — it’s an AI, and it gets better with your feedback, but it isn’t infallible. That’s why every idea shows a confidence score and reasoning: it tells you how sure it is and what would raise its certainty. Use the evidence, read the original conversations, and do your own sanity checks before building.
They’re validated in the sense that matters most: real people are asking for them. Each idea is traced back to actual demand signals from the web, and its evidence is shown alongside it. What you get is a researched starting point — the demand is already spotted, but the idea still needs to work for you.
Because the best business format depends on the problem. blink ideas picks the right venture type for each problem — a SaaS tool, a physical product, a local service, a marketplace, an agency, a subscription, or any other format — and only defaults to software when it’s genuinely the best fit.
Yes — that’s the whole point. The ideas are real problems people are asking to be solved. Pick one, follow the first validation steps, and make it yours.
Every idea is scored like an angel investor would weigh a deal, across seven weighted dimensions:
The score is a weighted sum from 0–100, shown as a pill with a bar next to it in the list. Green (70+) means a strong opportunity; yellow means promising; red means weak or unproven. Every idea also shows written reasoning for its score.
They answer two different questions. The score is how good the business looks on paper — demand, market size, scalability. The confidence score is how sure blink ideas is about its own judgment: how strong the evidence is, how familiar the market is, and how much uncertainty remains.
You can have a high-scoring idea with low confidence (big opportunity, but thin evidence) — the risky bet worth investigating. Or a decent score with high confidence (small but well-proven opportunity) — the safe bet. The confidence reasoning tells you exactly what would raise it.
The Score and Confidence filters narrow the list to your thresholds — type and they apply live. You can also sort by Most upvotes or Most downvotes to see what the community is most excited about (or most skeptical of), or search any keyword. The columns sort by clicking their headers.
This is the heart of blink ideas. Every comment you leave goes through a learning loop:
That’s why the ideas you see keep getting more refined, validated, and proven over time. Your voice literally shapes what the next generation of ideas looks like.
Your feedback is read, generalized into market lessons, and the specifics (brand names, one-off prices, idea-specific feature requests) are dropped. Vague praise or noise is skipped entirely. The goal is the underlying pattern, not the individual comment.
Votes are the community’s real-time signal — they rank ideas by what people are most excited about (or most skeptical of) in the sort-by-votes views. Written feedback is what feeds the learning loop that refines future idea generation and scoring: each comment is distilled into general market lessons that shape what blink ideas looks for next.
The chat knows everything about the product. Ask it for the latest ideas, search for an idea by keyword, get details on any opportunity, check the current stats, see what the worker is doing right now, or ask what the community feedback has taught it. It keeps your conversation in context across turns.
No accounts, no emails, no tracking. Votes, feedback, and chat sessions are tied to an anonymous token derived from your connection — so you get a consistent experience without us knowing who you are.
Not yet — but it’s on the roadmap. Until then, the fastest way to shape blink ideas is through your votes and feedback on existing ideas, which are distilled into lessons that guide every future idea generated.
It’s free today with no limits, and the plan is to keep the core experience free. If paid tiers ever arrive, they’ll be clearly announced in advance and the core discovery experience will stay free.
blink ideas is built by Drew Gallagher, a full-stack engineer with 7+ years shipping products. The long-term mission is simple: create more successful business owners. Questions or feedback? Send a note on LinkedIn.
No. blink ideas runs its own AI — you don’t need to configure anything. Just browse.
Skip the guesswork — the research is done, the demand is validated, the risk is reduced. All that’s left is picking your idea.
Browse Ideas →