How to find a startup idea if you don't know where to start

TrustMRR — see which products are already making money
TrustMRR is a database of startups where each project lists revenue verified through payment system APIs. All figures are verified directly via Stripe, LemonSqueezy, or Polar — without self-reported data or screenshots.

Visit the site, filter by niche or category — AI, SaaS, Developer Tools — and see which products are actually generating revenue: their monthly income, growth trends, and sometimes information about sales. This allows you to stop guessing whether there is demand in a niche and instead look at concrete numbers. Next, analyze the product's mechanics: what exactly it solves, who it sells to, and how much it charges. Use this as a basis to determine whether you can enter the same niche with a different positioning, lower price, or targeting a different audience.

Google Trends — check if interest in a problem is growing
Google Trends shows how the frequency of search queries on a topic has changed over months and years. This is useful not for finding an idea from scratch, but for testing a hypothesis: are people actually searching for a solution to this problem, and is their number increasing or decreasing?

Search for specific terms. The query "accounting software" will show a broad picture — it's unclear who is searching and why. The query "accounting software for sole proprietors without an accountant" already shows a specific person with a specific need. If such a query is consistently growing, it means more people have this problem and are actively searching for a solution.

IdeasAI — see which ideas people vote for
On ideasai.com, AI generates startup ideas, and users vote for those they find useful. Ideas with the most votes rise to the top — creating something like a ranking of real demand.

This is useful at the stage when you haven't yet decided on a niche. Instead of guessing what the market needs, you can see what people have already voted for. If an idea gathers thousands of votes, it means the pain point is recognizable and widespread. If there are few votes, the problem is likely either niche or irrelevant.

Personal experience — note what you do manually
Look at your current work and list tasks that repeat weekly, take up time, but don't require real thinking: data collection, report preparation, template-based correspondence, manual sorting. If this takes you several hours, chances are the same problem exists for people in similar roles at other companies.

This is how products that people actually buy emerge: not from abstract ideas, but from specific pains that founders themselves experienced daily.

Grok on X — search for complaints and requests directly
Grok is an AI assistant within the X platform, which has access to all public posts in real time. This makes it a convenient tool for finding ideas: you can ask a specific question and get a summary of what people are writing right now.

For example, ask: "What most annoys marketers in their work?" or "What tasks do founders want to automate?" On X, people often describe their workflows, share what isn't working, and openly talk about which tools they lack. Such posts are ready-made technical specifications for a product.

Reddit — read what people complain about in your niche
Reddit is thousands of communities on any topic, where people openly discuss their problems. Unlike other platforms, here people don't sell or pitch — they just talk about what isn't working.

People write about which tools they lack, what annoys them about current solutions, and what they've already paid for but remain dissatisfied with. If the same problem keeps coming up again and again, it's a signal.