Lovable ran out of credits and the app is not done: top up or hire someone?
Short answer: Top up if the app is moving forward and the remaining work is clear. Pay a developer if the last 20 credits or so bought you nothing, or the same bug keeps coming back. Lovable is a tool where you describe an app in chat and an AI writes the code, and it charges in credits. Top-up credits cost $0.30 each on the Pro plan (as of November 2026), so the money is rarely the problem. The problem is paying again and again for work that goes in circles.
Your credits are gone, the app is still not finished, and you need to know whether more credits will get it over the line or just burn more money.
I'm Ranjit, and I run Blur Studio: 15+ years building software, 120 apps taken live, and clients who have gone on to raise $10M+. I build AI systems and processes that help businesses do a lot more with a lot less time. I also work differently from most of what else is out there, more on that further down this page. Say hello.
Short version: the question is not "how many credits do I need?" It is "are the credits I spend still producing progress?" If yes, top up. If no, stop spending and change the plan. This page gives you a rule for deciding, a list of things that waste credits, and what happens to your app while you work it out.
What happens when your credits run out
Lovable uses one pool of credits for four things: building the app, chatting with Lovable, hosting the app and its built-in backend (called Cloud), and any AI features your finished app uses. When the pool is empty, per Lovable's own docs:
- Building stops. A new message opens a dialog asking you to top up or upgrade instead of building.
- AI features inside your published app stop working.
- Apps that use the built-in backend pause.
- Your published pages keep serving, and chat keeps working while your daily chat allowance lasts.
That third point matters if real customers use the app. Running out is not only a "my building is blocked" problem. It can switch off parts of a live product. If that is your situation, top up enough to keep the live app running first, then decide about the rest.
What refills, what does not
This is where most people get confused, so here it is in one table. All details are from Lovable's docs and pricing page as of November 2026.
| Credit type | Refills? | Expires |
|---|---|---|
| Daily build credits (5 a day, free plan capped at 30 a month) | Yes, every day at 00:00 UTC | End of each day, no rollover |
| Monthly plan credits (Pro starts at 100 a month for $25) | Yes, each billing period | 2 months after issue on monthly plans |
| Top-up credits (Pro and Business only) | No, you buy them | 12 months after purchase |
| Referral and bonus credits | When granted | Referral credits last 12 months |
Two facts people miss. First, the free plan's 5 daily credits are a small trickle: 30 a month at most, which is a handful of real changes. Second, credits are not refundable, so do not buy a big batch hoping to "see if it works".
You can top up in two ways: a one-time purchase, or auto top-up, which buys more when your balance drops below a number you set. If you use auto top-up, set a monthly limit. An app stuck in a loop plus auto top-up with no limit is how a $25 hobby turns into a bad month.
The break-even rule: top up or hire
Here is the rule I use. Look at your last 20 credits of spending (Lovable shows the cost of each message under "Credits used" in the message menu) and ask three questions:
- Did the app get measurably closer to done? Not "different". Closer. One more working screen, one bug truly gone.
- Is the same problem back for the third time? If you have asked Lovable to fix the same bug three times and it keeps returning, more credits will not fix it. The AI is guessing at a problem it cannot see.
- Can you list what is left on one page? If you can write down the remaining jobs in plain words, the work is bounded. If every fix creates two new problems, it is not.
Three yeses: top up. Any no: stop and get a second opinion before buying credits.
A worked example (made-up figures)
Say you built a quote-request app. You have used 300 credits over a month. The last 40 credits went on trying to fix the login, which still fails for some users. Here is the comparison, using Lovable's Pro top-up price of $0.30 per credit:
- Keep going yourself: another 40 credits is $12 in top-ups. That sounds cheap. But if you spend 8 more hours at a value of $50 an hour, the real cost is $412, with no guarantee the login gets fixed.
- Pay someone to finish: say a developer quotes a fixed $2,000 to finish it (a made-up figure, get real quotes), covering login, testing and putting the app live. You pay once and the problem is gone.
So the break-even is not credits against dollars. It is your hours plus credits against a fixed price. If the remaining list is short and clear (three or four small changes), top up and finish. If the list is long, or has anything in the risky categories below, a fixed price is usually cheaper once you count your own time.
What burns credits for no gain
These are the habits that eat a credit balance without moving the app forward.
- Repeating "fix it" on the same bug. Each retry costs credits. After two failed tries, change the approach: describe the exact steps that trigger the problem, paste the error text, or ask a question instead of asking for a change.
- Using Build mode to think. Lovable's Plan mode costs 1 credit per message and never changes your code. Chat mode messages are priced at a fraction of a credit and draw on a daily chat allowance first (Lovable describes its chat pricing as a temporary offering that can change, so check the current terms). Use these to ask "why is this failing?" before you pay for a build.
- Big vague prompts. "Make the whole app better" triggers a large, costly rewrite. Small, specific requests cost less. Lovable's own examples show a button color change at about 0.5 credits and a landing page with images at about 2.
- Generating images and video in the builder. These count as build usage. Make or buy your images separately and upload them.
- Design fiddling. Moving a button five times costs five messages. Decide the layout on paper first.
- Letting a long run keep going. Lovable pauses a long message at a check-in level (20 credits by default) so you can continue or finish up. Do not just keep pressing continue.
To see where credits went, open Workspace settings, then Plans and credit usage, then Usage details. You can filter by project and by person. If you want a safety net, paid plans have Usage limits and alerts, which let you cap spending per project or workspace and warn you at a balance you choose.
Problems credits usually will not fix
Some jobs are poor fits for chat-driven building, however many credits you throw at them:
- Payments. Money handling needs careful testing in both test and live modes.
- Login and permissions. If users can see each other's data, that is a security fault, not a styling bug. See my guide on why AI-built apps break.
- Anything that touches live customer data. A mistake here is expensive to reverse.
- Odd bugs that only appear for some users. The AI cannot see what those users see.
If your problem is in this list, a person who can read the code and the logs will be faster than any number of credits. My piece on whether AI can finish a half-built app or you need a developer goes deeper on that line.
Your options, in order of cost
- Wait for the daily credits. Free, but only 5 a day. Fine for tiny edits, useless for a rescue.
- Top up a small amount and use Plan mode. Spend 5 to 10 credits understanding the problem, not changing code.
- Upgrade the plan. Lovable's own docs say a higher plan or credit tier is usually better value than repeated top-ups if you need more every month.
- Get a developer to review and finish. See how to hire a Lovable developer for what to ask and what a fair job looks like.
- Move to a different builder. Sometimes another tool suits your app better. My Lovable alternatives guide compares them. Note that other builders have the same credit trap: my article on why Emergent credits run out fast (Emergent is another AI app builder) shows the pattern.
Getting your code out so a developer can help
You do not need credits to hand the project to someone. Lovable has Git sync, which connects your project to a GitHub, GitLab or Bitbucket repository you own (GitHub is a website that stores code and its full history). Per the docs, Git sync is available on all plans, including free, and changes sync both ways, so a developer can work in their own tools and the changes appear back in Lovable. Paid plans can also download the codebase from the code editor. Lovable's pricing page says you own your code.
My guide on how to export code from an AI app builder walks through the steps and what does and does not come with the code. One caution: the database and login that Lovable Cloud runs for you are separate from the code. Ask the developer how they will handle those before anyone moves anything.
What to give a developer so the quote is fair
- A link to the repository, or a zip of the code.
- A one-page list of what works, what is broken and what is missing.
- The steps that trigger each bug, with screenshots.
- Any error messages you have pasted into chat before.
- Logins for the services involved (payments, email, database), shared securely, not by email.
Ask for a fixed price in writing, a clear definition of "done", and a short review first. A good developer will look before quoting.
How I'd approach this
I would not buy more credits tonight. I would spend ten minutes on the three questions above, then write the one-page list of what is left. If the list is short and clear, I would top up a modest amount, use Plan mode first, and set a hard monthly limit so nothing runs away.
If the list is long, or login, payments or customer data are in it, I would stop paying credits and get the code reviewed by a person. That is what a Blur Studio finish is for: I look at what the AI built, tell you plainly what is wrong, and fix it for a fixed price. My fixes and finishes start around $1,500. I would rather tell you a top-up will do the job than sell you a finish you do not need.
Either way, copy your code to a repository you own now. It costs nothing and it means you are never stuck again.
Questions people ask
Do Lovable credits come back on their own?
Only some do. The free daily build credits (5 a day, up to 30 a month on the free plan) refresh every day at 00:00 UTC, and paid plans add monthly credits each billing period. Top-up credits do not refresh, you buy them. Check your Credit balance dialog in Workspace settings for the exact expiry of each type.
How much does a Lovable top-up cost?
As of November 2026, Lovable's docs list Pro top-ups at $15 for 50 credits ($0.30 each) and Business top-ups at $30 for 50 credits. Top-ups are only for Pro and Business plans, and top-up credits last 12 months. Check Lovable's pricing page for the current price before you buy.
Will my published app go down if my credits run out?
Not completely. Lovable's docs say your published pages keep serving. But AI features inside your app stop working and apps using the built-in backend (Cloud) pause. If customers use those parts, top up enough to keep them running before you decide anything else.
Are Lovable credits refundable?
No. Lovable's pricing page says credits are not refundable or redeemable for cash. If you cancel, you can keep using remaining credits until the end of the billing period. That is why I suggest buying small amounts until you know the app is on track.
Can I keep building outside Lovable?
Yes. Git sync connects your project to a GitHub, GitLab or Bitbucket repository you own, on all plans, and a developer can work on it elsewhere. See how to export code from an AI app builder for the steps and the things that do not move with the code.
How do I stop spending credits by accident?
On paid plans, open Workspace settings, then Usage limits and alerts. You can cap how many credits a person, a project or the whole workspace can use, and set a low-balance alert. If you use auto top-up, set a monthly limit too. Use Plan mode to ask questions, since it never changes your code.
When is a developer cheaper than more credits?
Usually when the remaining work includes logins, payments or customer data, or when the same bug has come back three times. Count your own hours as well as credits. My guide to hiring a Lovable developer covers what a fair job looks like.
More in this topic: Problem solving: AI-built and no-code apps (26 guides) →