How to get your first 100 users for an app you built with Lovable, Cursor or Bolt

You prompted your way to a working app. It has a domain, a login and a Stripe button. What it does not have is users, and the people who make AI builders never tell you that this part is slower than building was.

The first 100 users almost never come from one launch. They come from a few dozen small, specific actions: one Reddit reply here, one Show HN there, one DM to a person who complained about the exact problem you solved. This is the plan, in the order that works.

Day 1: make the landing page answer three questions

Before anyone sees the app, the page has to answer "what is it", "who is it for" and "what do I do now" in the first screen. Most AI-generated landing pages fail this: they open with a slogan, three feature cards and a gradient, and the visitor still cannot tell what the product does.

1. Headline: the outcome, in plain words. "Turn Notion pages into a client portal" beats "Reimagine collaboration". If a stranger cannot repeat what your app does after reading the headline once, rewrite it.
2. Subhead: who it is for and what it replaces. Name the audience and the thing they do today. "For freelance designers who send Google Drive links to clients."
3. One button. Sign up, start free, or book a call. Not three.
4. Proof, even if small. A 20-second screen recording (Loom or QuickTime) of the app doing the thing. If you have three users, quote one of them by name with permission.
5. Make it crawlable. A real <title>, a meta description, an Open Graph image, and a sitemap. Run it through our free site grader or Google's Rich Results Test; both take a minute. Lovable and Bolt projects often ship with the default title and no OG image, which makes every link you share look blank.

Install analytics before you post anywhere. PostHog and Plausible both have free tiers and a one-line script. You need to know which of the actions below actually brought people, or you will keep doing the ones that felt good instead of the ones that worked.

Day 2: find the ten threads where people already ask for it

Somewhere on Reddit, Hacker News, X or a niche forum, people are asking for what you built. They said it in their own words, which are rarely your words. Your job is to find those threads and answer in them. This is the highest-return hour in the whole plan.

How to search:

• On Google: site:reddit.com "is there a tool" [your problem], site:reddit.com "recommend" [category], site:news.ycombinator.com "Ask HN" [category]. Use the "Past year" filter.
• On Reddit itself: search the phrase your customer uses, sort by New, and read the comments. The best threads are questions with two or three answers and no clear winner.
• On X: search "anyone know" [category] and "looking for a" [category], filter to Latest.
• Set up alerts for the future. F5Bot emails you whenever a keyword appears on Reddit or Hacker News, and it is free. Google Alerts covers the wider web.

Build a list of ten threads. Then answer each one properly: solve the person's problem first, in two or three sentences, and mention your app in one sentence at the end with a disclosure ("I built this, so take that into account"). Threads that are a year old still get search traffic; a good answer on an old thread keeps sending you visitors for months.

Do this by hand, from your own account. Do not automate it. Reddit's spam filters and moderators are good at spotting accounts whose only activity is dropping the same link, and a ban is permanent for that account. AutoTraction, which is on a waitlist for a September 2026 launch, drafts these replies for you and hands them over to paste for exactly this reason.

Day 3: post it where posting is the point

A few places exist specifically for showing what you built. They are small, but they are full of people who try new things and tell others. Post to these in this order, one per day, so you can answer comments properly.

• Lovable Launched (launched.lovable.dev) if you built with Lovable. It is a directory of Lovable apps and the audience is other builders, which is useful if your product serves them.
• r/SideProject. Self-promotion is the point of the subreddit, but the moderators remove bare links, waitlist pages and reposts. Show the working product, say what you built it with and why, and reply to every comment.
• Show HN on Hacker News, only if people can try the app without signing up. The Show HN guidelines exclude landing pages and sign-up pages, and asking friends to upvote gets the post killed. Title format: "Show HN: [Name], [what it does in plain words]". Post a first comment explaining why you built it and what you are unsure about.
• Peerlist Launchpad, Uneed, Tiny Startups and MicroLaunch. Smaller launch boards that were all active as of mid-2026. Each brings a handful of visitors; the value is mostly the backlink and the habit of writing a tight description.
• Product Hunt, later. Since the Featured/All split, only launches the editorial team picks as Featured appear on the homepage and in the app. Unfeatured launches get very little. Wait until you have a polished product and a few users who will comment, then launch on a weekday.

Our full list with the rules for each is in Where to post your side project in 2026.

Day 4: message twenty people, one at a time

Cold outreach works at this stage because it is personal. Find twenty people who have publicly described the problem: a tweet complaining about it, a LinkedIn post, a Reddit comment, a review of a competitor. Message each one with three sentences:

Saw your post about [the problem]. I built [name], which does [the outcome] for [the audience]. If you want to try it, here is a link and I will personally set it up with you.

Offer a call. At this stage a 15-minute call with one user teaches you more than a hundred sign-ups you never talk to. Use Cal.com or Calendly for the link. Track who you contacted in a spreadsheet, because you will forget.

Day 5 to 7: reply, follow up, and write one honest post

Everything you posted is now generating comments. Answer all of them the same day. The quality of your replies is what people judge, not the launch post.

Follow up once with everyone who signed up but did not come back. A plain email, from you, asking what stopped them. Half will not answer; the ones who do will tell you what to fix next.

Then write one post about what happened, on X or LinkedIn or your own blog: what you built, what you tried, what worked, the numbers. Founders who build in public get their next hundred users from the people who watched them get the first hundred. Keep it factual. "Posted to five places, 212 visitors, 19 sign-ups, 4 people on calls" is more interesting than any launch announcement.

What to skip

• Paid ads. You do not know your conversion rate or your pitch yet. Ads amplify whatever you already have, and right now you have a guess.
• Posting in every subreddit at once. Cross-posting the same link across ten subreddits in an hour is the most common way new accounts get banned. Read the rules by subreddit first.
• SEO content. Articles take months to rank. Write them, but do not expect them to bring your first users.
• Influencers and affiliates. Nobody promotes a product with no users. Come back to this at 1,000.

The routine after the first week

The first 100 users usually take a few weeks of this, not one. Once the burst is over, what matters is showing up daily: check your alerts, answer one or two threads, message a couple of people, and post one thing. That is about four minutes a day if the searching and drafting are done for you, which is the job we are building AutoTraction to do. Until then, do it by hand and keep the spreadsheet. The spreadsheet is the marketing plan.