But nobody else can build such a fine web container so we can have a development server run inside a Chrome browser.
Why Bolt.new and Similar AI Coders Freeze or Stall During Code Assembly
Several user reports and technical threads highlight that Bolt.new and related StackBlitz AI coding frameworks often freeze or stall during code assembly due to a combination of high resource usage, memory leaks, and inefficient code regeneration processes:
- Memory Leaks and High Resource Usage: Users frequently encounter out-of-memory (OOM) errors, especially when running Bolt.new in Chrome. Memory usage can spike rapidly, leading to browser crashes or unresponsive sessions. This issue is sometimes less severe in alternative browsers like Edge, suggesting browser-specific memory handling differences714.
- Glitchy Streaming and Build Processes: The streaming response for code generation is described as “glitchy and irregular,” with slow performance and warnings about resource thresholds being approached (e.g., memory usage at 75%, CPU load at 60%). This can cause the interface to freeze or stall, particularly during complex or lengthy code assembly13.
- Full App Recoding on Minor Changes: Bolt.new tends to regenerate the entire application codebase even for small edits, which is highly inefficient and can lead to repeated stalls and excessive resource consumption6.
How to Reduce Token Consumption When Bolt.new Stalls or Makes Errors
- Minimize Prompt Length and Complexity: Avoid overly long or complex prompts, as exceeding supported prompt sizes can cause the system to become unresponsive or waste tokens on incomplete attempts46.
- Request Incremental Changes: Make small, incremental requests rather than asking for large-scale refactors. This reduces the chance that Bolt will attempt to rewrite the entire codebase, which is a major source of token waste6.
- Use Debugging Techniques: Ask Bolt to add extensive logging to help debug issues before requesting fixes. This can help isolate problems and avoid repeated, token-consuming trial-and-error cycles8.
- Manual Intervention: When Bolt stalls or produces errors, consider manually resolving the issue or pasting error messages into another LLM (like Claude, ChatGPT, or Perplexity) for targeted troubleshooting9.
Common User-Reported Issues Causing Freezing and Token Waste
- Infinite Update Loops: Bolt sometimes takes “an infinite amount of time and ultimately produces nothing,” especially when updates trigger full app regeneration26.
- Rate Limits and Token Exhaustion: Free accounts are particularly susceptible to hitting rate limits after repeated errors or incomplete code generations, leading to wasted tokens and stalled progress36.
- Unreliable Code Changes: Even minor requests (such as adding a filter) can trigger a cascade of errors, requiring multiple retries and consuming millions of tokens in the process610.
- Project Disappearance and Unresponsive UI: Users report projects vanishing or the UI becoming unresponsive, necessitating a browser or session restart4.
Known Solutions for Memory Crashes and Slow Streaming Performance
- Close Unnecessary Tabs and Applications: Free up system memory by shutting down other browser tabs and background processes before running Bolt.new45.
- Use a Chromium-Based Browser (Preferably Edge): Some users report better memory handling and fewer crashes on Edge compared to Chrome74.
- Disable Browser Extensions: Selectively turn off ad blockers and other extensions, which can interfere with Bolt’s WebContainer runtime4.
- Optimize Build Settings and Code: Use efficient build tools (like Vite), minimize dependencies, and implement code splitting to reduce memory load and build times5.
- Monitor and Manage Memory Usage: Use Chrome DevTools or similar tools to track memory consumption and identify leaks5.
- Split Large Projects: If your project is large, break it into smaller modules or components to avoid exceeding memory and prompt limits4.
Typical Troubleshooting Steps for Repeated Code Errors and Incomplete Files
- Check and Correct Build Settings: Ensure your build configuration is correct and that you are using the appropriate commands for deployment and testing5.
- Manual Debugging and Logging: Add logging statements to the code to pinpoint issues, then ask Bolt to debug based on those logs8.
- Leverage External LLMs for Error Analysis: Copy error messages into other AI tools for more context and alternative solutions9.
- Restart Browser or Switch Browsers: If Bolt freezes, restarting the browser or switching to another Chromium-based browser can sometimes resolve the issue74.
- Contact Support for Serious Bugs: If projects disappear or persistent errors occur, reach out to Bolt.new support for assistance4.
Summary Table: Common Issues and Solutions
| Issue | User-Reported Cause | Practical Solutions |
|---|---|---|
| Freezing/Stalling | Memory leaks, high resource usage, full app recoding | Close tabs, use Edge, split projects |
| Excessive Token Consumption | Full app regeneration on minor edits | Incremental requests, debug first |
| Out-of-Memory (OOM) Errors | Large dependencies, inefficient code | Optimize code, limit dependencies |
| Rate Limits and UI Freezes | Repeated errors, incomplete generations | Manual intervention, contact support |
These strategies can help reduce frustration, token waste, and downtime when using Bolt.new and similar browser-based AI coding tools246789.
StackBlitz’s Recognition of Bolt.new Issues and Their Response
StackBlitz is aware of several user-reported challenges with Bolt.new, particularly regarding performance, error handling, and resource usage. While the company celebrates Bolt.new’s rapid adoption and transformative impact on web development, user feedback and independent reviews make it clear that issues such as code overwriting, lack of chat history, version control limitations, and high token consumption due to AI-induced errors have been noted by the community357.
Evidence of Issue Recognition
- User Feedback and Public Discussions: Users have voiced concerns on platforms like GitHub and in product reviews, citing slowdowns, excessive token consumption, and performance degradation after upgrading to paid plans57. These public threads and reviews indicate StackBlitz is aware of the pain points, as users are directly reporting them in official channels.
- Iterative Product Development: StackBlitz has a history of rapidly iterating on its products in response to user needs and feedback, as highlighted by the company’s pivot to AI-powered development and the integration of features like mobile app support and Figma conversions36. This suggests a willingness to address shortcomings as they arise.
Actions and Solutions in Progress
- Ongoing Improvements: StackBlitz continues to “iterate and improve Bolt.new,” with a clear understanding that resolving performance and usability issues is crucial for maintaining user trust and long-term growth3. The company’s roadmap appears to prioritize stability and user experience enhancements.
- Usage-Based Pricing Model: In response to concerns about cost and token waste, StackBlitz shifted to a usage-based pricing model, aiming to better align user expenses with actual output and reduce frustration over wasted tokens due to AI errors36. This model is intended to give users more control over costs and to incentivize more efficient AI usage.
- Technology Enhancements: The underlying WebContainers technology is regularly updated to improve browser-based performance, dependency management, and error handling16. StackBlitz’s integration with third-party services (like Supabase and Netlify) is also designed to reduce error rates and streamline deployments6.
- Community Engagement: By maintaining open-source repositories and public issue trackers, StackBlitz invites direct user feedback and bug reports, which can be acted upon in future updates17.
Remaining Challenges
Despite these efforts, some users continue to experience slowdowns and performance issues, especially with complex projects or after upgrading to premium plans57. The company acknowledges that addressing these technical challenges is an ongoing process and is critical for sustaining its rapid growth and positive reputation3.
In summary: StackBlitz recognizes the issues affecting Bolt.new, as evidenced by user feedback and public discussions. The company is actively working to address these challenges through product iteration, pricing adjustments, technology upgrades, and community engagement, but some problems—particularly those related to performance and resource usage—remain a work in progress357.
Leave a Reply