This article contains affiliate links. If you purchase through these links, we may earn a commission at no extra cost to you.
Leonardo AI goes down, stalls, or throws cryptic errors more often than you’d expect for a platform this polished. The good news? 90% of issues have a fix that takes under two minutes.
Check the status page first. Every time. Go to status.leonardo.ai before you do anything else. I’ve spent embarrassing amounts of time troubleshooting what turned out to be a scheduled maintenance window.
If the status page shows all-green and you’re still stuck – keep reading.
Login and Account Access Problems
Can’t get past the front door? This one’s frustrating because you haven’t even started yet.
OAuth login loops are the most common culprit. If you’re using Google or Apple sign-in and getting bounced back to the login screen over and over, the fix is usually one of these:
- Clear your browser cookies for leonardo.ai specifically (not all cookies – just theirs)
- Try a different browser entirely
- Disable any browser extensions that modify headers or block trackers – Privacy Badger and uBlock Origin can interfere with OAuth flows
Email verification not arriving is annoying but simple. Check spam, obviously. If it’s not there after five minutes, hit “Resend verification” – but wait the full five minutes first, because their mail queue sometimes lags. Still nothing? Try adding [email protected] to your contacts and requesting again.
Password reset not working – if you originally signed up with Google OAuth, you don’t have a password to reset. Log in with Google. I’ve seen people confused about this more times than I can count.
Image Generation Not Working or Stuck
This is the big one. You hit generate, the spinner appears, and then… nothing. Or it finishes and shows an error. Or it generates something completely black or blank.
Queue overload. During peak hours (generally 9am-1pm EST and 7pm-10pm EST based on my usage patterns), generation queues can back up significantly. If your generation is sitting at “pending” for more than three minutes, you’re stuck in the queue. Options: wait it out, switch to a less popular model, or reduce your resolution. The Phoenix model tends to have shorter queues than SDXL variants during peak times.
Insufficient credits. Leonardo uses a token/credit system that’s – honestly – kind of complicated. You might have subscription credits, free daily tokens, and model-specific costs all interacting. Check your token balance in the top right corner. If it shows 0, you’re out regardless of subscription status.
One gotcha: some models cost more tokens than others. If you’ve been using Alchemy mode or upscaling, you’re burning tokens faster than you think.
NSFW filter triggers. The content filter isn’t subtle – it just fails the generation with a generic error. If you’re generating fantasy or figurative art and hitting this repeatedly, try rephrasing. Avoid anatomical terms even in clinical contexts. The filter is keyword-heavy, not context-aware.
Blank or black output usually means one of two things: the guidance scale is too high (try dropping it from 7 to 3-5 for most models), or there’s a conflict between negative prompt and positive prompt where you’ve essentially told the model to generate nothing.
Credits Not Being Applied or Subscription Issues
You subscribed. It charged your card. You still can’t generate anything. This one’s infuriating.
Billing sync lag is real – I’ve seen it take up to 20 minutes after a subscription purchase before tokens appear in an account. Close the tab, wait 15-20 minutes, log back in fresh. Don’t keep refreshing.
Daily token reset timing. Free tokens reset at midnight UTC. Not your local midnight. If you’re in the US, that’s 7pm or 8pm EST depending on daylight saving time. Worth knowing if you’re burning through your daily allotment and wondering when they’ll refresh.
Plan tier confusion. The Apprentice, Artisan, and Maestro plans have different monthly token totals AND different concurrent generation limits. If you’re on Artisan and generating multiple images simultaneously, you might hit the concurrency cap before running out of tokens. The subscription page shows this breakdown, but it’s not surfaced prominently in the main interface.
If you genuinely believe you were charged and tokens never appeared, hit up support. Keep a screenshot of your payment confirmation handy – it speeds things up considerably.
Slow Generation or Low Quality Output
Got it generating but not happy with the results?
Model selection matters more than anything else. The default model isn’t always the best model for your use case. For photorealistic images, Phoenix is currently the strongest option. For stylized illustration, the older SDXL-based models often produce better results than people expect. Spend an hour experimenting before assuming the platform can’t do what you need.
Resolution and quality settings. Higher resolution takes longer – obviously – but also tends to produce sharper detail. The tradeoff is tokens. If speed matters more than quality, drop to 512x512 and see if you like what you get before committing to a full 1024x1024 generation.
Guidance scale and steps. These two settings are where most beginners leave quality on the table. Guidance scale (sometimes called CFG scale) controls how strictly the model follows your prompt. Too low (under 3) and you get creative but unpredictable output. Too high (over 10) and the image gets oversaturated and weird. Steps control generation quality – 30 is usually plenty, and going above 50 rarely produces meaningful improvement while costing time.
API Not Responding
For developers using the Leonardo API, a few things to check when calls are failing:
Key regeneration. API keys can expire or get rotated without warning if there’s been suspicious activity on your account. Regenerate your key in the platform settings, update your code, and test again before assuming the API endpoint has changed.
Rate limits. The free API tier is quite constrained – I think it’s 150 requests per day on the base plan. If you’re hitting 429 errors, that’s your rate limit. The limit resets at midnight UTC, same as tokens.
Endpoint changes. Leonardo has updated their API structure a few times. If you’re using code from a tutorial that’s more than six months old, check the current API documentation to make sure the endpoint and request format haven’t changed. The v1/generations endpoint specifically has seen a few parameter updates.
CORS issues when calling from a browser – don’t. The Leonardo API isn’t meant to be called directly from frontend JavaScript. Run it server-side and you’ll save yourself the headache.
When to Actually Contact Support
Before you open a support ticket, try the Discord first. The #bug-reports and #help channels are surprisingly active, and you’ll often find someone who hit the same issue an hour ago with a working fix. Discord is at discord.gg/leonardoai.
Contact support directly when:
- You’ve been charged but tokens never appeared after 30+ minutes
- Your account is locked out and all standard recovery steps have failed
- You’re seeing persistent API errors that started after no code changes on your end
- Content was incorrectly flagged and removed
The help center is at leonardo.ai/help. Response times vary – expect 1-2 business days via ticket.
Frequently Asked Questions
Why does Leonardo AI say “generation failed” with no other message?
Usually a content filter trigger or a server-side error during high traffic. Check the status page first. If the platform’s healthy, try simplifying your prompt – remove any words that might trigger content filters, reduce the resolution, and disable Alchemy mode temporarily. If it still fails with a simple prompt, wait 30 minutes and try again.
Why are my Leonardo AI tokens not refreshing?
Free daily tokens reset at midnight UTC – not your local time. If you’re in North America, that’s late afternoon or evening. Paid subscription tokens reset monthly on your billing date. If neither applies and your tokens aren’t refreshing, log out and back in to force a session refresh, or try clearing your browser cache.
Leonardo AI is loading but images aren’t displaying – what’s wrong?
Likely a CDN or browser caching issue. Hard refresh the page (Ctrl+Shift+R on Windows, Cmd+Shift+R on Mac). If images still won’t load, try an incognito window or a different browser. Occasionally their image delivery CDN has regional issues that don’t show up on the main status page.
Why is Leonardo AI so slow today?
Check the status page and their Discord. Peak usage hours see slower generation times across the board. Switching models and reducing resolution are the fastest ways to get speed back. The platform also runs slower on weekends when hobbyist traffic spikes.
Can I use Leonardo AI without an account?
No. You need an account to generate images, even on the free tier. The free tier does give you a reasonable number of daily tokens to experiment with – it’s worth creating an account even if you’re just testing the platform before committing.
If Leonardo AI’s limitations are getting to you, it’s worth knowing what else is out there. Our best AI image generators roundup for 2026 compares the full field, and our Leonardo AI review covers what it does well (and where it falls short). If you’ve been hitting the content filter repeatedly, Midjourney handles figurative art with fewer restrictions. And if Ideogram’s been giving you grief too, the Ideogram troubleshooting guide follows the same diagnosis-first approach.
Most Leonardo AI problems are temporary. The platform’s solid when it’s running right – and usually it runs right.