Questions · Published
How much does it cost to build an app with Lovable, Bolt or Replit?
TL;DR
A simple app can be built in an afternoon for little or nothing: a free plan or a $20 to $30 monthly plan usually covers it. A product you can sell, with accounts, payments and teams, costs more: founders report spending anything from under $150 to several thousand dollars, and most of the difference comes from fixing and rebuilding. Once it's live, a small app usually costs under $100 a month to run.
Lovable, Bolt, Replit and v0 are AI app builders. You describe what you want in plain English, and they write the code, set up a database and sign-in, and put the app online. You pay a monthly plan that includes a set amount of building, measured in credits or tokens, and buy more when it runs out.
What it costs to build
A simple app, like a booking page for a dog walker, can be built in an afternoon without using up the building included in a $20 to $30 monthly plan. A product you can sell needs more: sign-up, teams, payments, settings and the fixes each of those brings. Founders building that kind of product often spend a few hundred dollars, and some spend thousands.
Most of that money goes on fixing things. You ask for a change, something else breaks, you ask the builder to fix it, and every attempt uses credits. Changes also get more expensive as the app grows, because the builder reads more of your project each time, which Bolt's guidance on saving tokens points out.
Founders who post about their spend show how wide the range is. On Replit, one reports building several apps for under $150 each, writing a business plan and requirements before each one. Another spent $2,500 on one feature-rich product, and another about $4,100 across five products over 13 months. These are self-reported and the apps differ, but they show how far the cost can stretch.
What the plans cost
Each builder has a free plan for trying it out, and paid plans start at $20 to $30 a month. Prices as of October 2026:
| Builder | Free plan | Starting paid plan | What's included | Hosting |
|---|---|---|---|---|
| Lovable | Daily credits | Pro, $25/month | 100 credits | Included |
| Bolt | 1M tokens a month | Pro, $25/month | 10M tokens, unused tokens roll over a month | Included |
| Replit | Limited | Core, $20/month | $20 of usage | Included, usage billed |
| v0 | 7 messages a day | Plus, $30/user/month | $30 of credits, plus $2 a day | Separate, on Vercel |
Each builder counts usage in its own unit, so the allowances can't be compared directly. On Lovable, for example, changing a button colour uses half a credit and adding sign-up and login uses a little over one. When the allowance runs out you buy more credits or upgrade, and that's where most building costs come from.
What it costs to run
Once it's live, a small app costs little to run. Lovable, Bolt and Replit include hosting, a database and sign-in in their plans. If you use separate services instead, such as Supabase for the database, Clerk for sign-in or Resend for sending emails, each has a free plan big enough for your first customers and paid plans from about $20 to $25 a month. Card payments cost a percentage of each sale, such as 1.5% plus 20p for a UK card with Stripe. For a small business app with tens of customers, under $100 a month is a sensible figure to plan with.
If your app uses AI itself, add model costs. Every customer request costs you something, so price your plans with that in mind.
How to spend less building
- Decide what you're building before you prompt. Know who it's for, what the first version has to do and what you'll charge. Bolt's guidance says planning first saves tokens, and a March 2026 study found coding agents built more faithful versions when they had the full requirements at the start.
- Keep the first version small. Every feature you add before anyone pays is one more thing to build, fix and pay for. See what the first version should have.
- Roll back when a fix goes in circles. Going back to the last working version and rewording your request usually costs less than asking for fix after fix.
A few hundred dollars on a version you can show five potential customers is money well spent. Thousands on features nobody has asked for is what to avoid, so it's worth working out whether to build yet before you start.