Urgent.News

What's breaking now, across thousands of outlets.

Tech

Why your Bolt.new or Cursor app works for you and breaks for everyone else

There is a moment in every Bolt.new or Cursor build where the app stops being a demo. Up to that point it is fun. You describe a screen, the thing appears, you describe a behaviour, the thing does it. The demo video writes itself. Then you let one real person use it, and you find the parts that only ever worked because you were the one using it: the seed data that only exists on your laptop, the…

Building an AI-generated application like Bolt.new or Cursor can seem straightforward at first. You describe a screen, an action happens, and the demo video writes itself. However, this illusion crumbles when a real person uses the app for the first time. The problems that only worked in the controlled environment of the developer's machine suddenly surface.

One of the most common issues is the lack of a database. Instead of pulling data from a live database, the app uses seed data hardcoded into the code. This means that adding a fourth item causes it to disappear, leaving a blank screen or an undefined error for the user. The empty state was never designed, making it crucial for new users to have some content from the start.

Another problem is the absence of a proper error handling mechanism. Third-party APIs, such as payments, email, maps, storage, AI inference, or authentication, can fail in various ways. Without a fallback, users might see a spinning spinner that never resolves, leaving them confused and frustrated. Similarly, if the code calls these services without considering failure scenarios, users will encounter a blank screen or a spinner.

Furthermore, the deployment process can be a significant hurdle. The app works only because it is deployed from the developer's machine. The environment variables live in a .env file on the developer's machine, the database is a local file, and the build step assumes a folder that only exists locally. This makes it impossible for other users to access the app, rendering any fixes moot.

To avoid these pitfalls, developers should focus on building a real application that runs somewhere other than the developer's machine. The deployment should include a URL that can be accessed by others, with connection strings stored in the host's environment and migrations run by the deploy step. Once the app is live, test it thoroughly with real data and ensure that it gracefully handles empty states and errors.

Building empty and error states is equally important. Instead of a spinning spinner, provide a sentence that clearly communicates the issue and what the user should do next. This could be a message like, "An error occurred. Please try again later." Additionally, handle third-party API calls with timeouts and fallbacks. For instance, if a payment provider is down, the app should display an error message instead of leaving the user staring at a blank screen.

Finally, ensure that the app is fully functional before considering it complete. Can a new user sign up, perform the primary action, and achieve the expected result without any assistance? If not, the app is still a prototype that needs further development. For those who need help troubleshooting specific issues, resources like AppStuck provide platform-specific guides for Bolt, Lovable, Replit, Cursor, Base44, and v0. The troubleshooting process is similar across all platforms, but the tool names may vary.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

How to Track WhatsApp Enquiries in Your CRM After a Website Click

Track a WhatsApp button click as website intent, an inbound message as a conversation, and sales qualification as a CRM outcome. Keep the three counts separate.

  • Clicking WhatsApp button triggers website action
  • First valid inbound message starts conversation
  • Track distinct events: click, conversation start, qualification

Six reasons your Lovable app breaks in production (and the order to fix them)

Every Lovable build breaks the same way, and it is almost never the interesting way. You spend a weekend making something real: a marketplace for dog walkers, a booking tool for a barbershop, an…

  • Database connectivity issues cause failures when real users access the app
  • Authentication problems arise from incorrect callback URLs and session cookie flags
  • Secrets in code and bundles must be moved to environment variables for security

Social Media Character Limits: What 11 Networks Actually Count

Social media character limits look like the simplest fact in marketing: one number per network. Read the developer documentation behind those numbers and they stop being simple.

  • Bluesky limits characters to 300 graphemes or 3,000 bytes
  • Threads counts emojis by UTF-8 byte count, not by visual count
  • Instagram limited hashtags to 5 since December 2025

A job board that agents pay for: four x402 bugs we hit before a directory would list it

We run a small job board for AI agents at agent-job-board.obtranzat.workers.dev . No accounts, no API keys: an agent pays 0.02 USDC to read the board, 0.50 to post a job, 0.01 for an aggregated feed…

  • Agents pay 0.02 USDC to read job board, 0.50 USDC to post job
  • x402 v2 rail on Base (eip155:8453) facilitates payment process
  • Multiple fixes addressed HEAD request oversight, discovery, multi-rail handling, error handling

Why We Switched From Homegrown Feature Flags to LaunchDarkly (And What We Learned)

We built our own feature flag system three years ago. It started as a simple Redis hash: feature:new-checkout -> {enabled: true, percentage: 50} . A Rake task to flip it.

  • Switched from homegrown to LaunchDarkly after three years
  • Homegrown system lacked audit trail, segmentation, kill switch
  • Migration took two sprints, impacted day-to-day operations positively

More from Tuesday 17 November →