How modelBridge Handles Errors
What you see when something goes wrong
Section titled “What you see when something goes wrong”When an error occurs during generation, modelBridge shows a banner above the Generate button with three things: what happened, why, and what to do next. Messages are written in plain language — you’ll never see raw API responses, HTTP status codes, or technical field names.
Each banner is color-coded by type:
- Red — something about your input needs fixing (image too small, video too long, wrong format)
- Amber — action required on your end (API key issue, balance empty, content flagged)
- Blue — temporary issue that will resolve on its own (server busy, network timeout)
Some banners include a “Read more” link that opens the relevant section in these docs. This link only appears when a verified documentation page exists for that specific error — you’ll never land on a broken or missing page.
If you see a generic message like “Something went wrong processing your request”, it means modelBridge caught an unexpected response from fal.ai. The error has been logged internally and will be addressed in a future update. You won’t be charged for failed generations.
Known vs unknown errors
Section titled “Known vs unknown errors”Known errors have specific messages tailored to the exact problem. For example, if your image is too small for a model, you’ll see the exact minimum dimensions required and a suggestion to use a larger image. These errors include fix steps and, where available, a “Read more” link to the relevant docs section.
Unknown errors are new error types that haven’t been documented yet. When this happens, modelBridge shows a safe, generic message — no raw technical output, no broken links, no confusing API details. The plugin handles it gracefully while the issue is reviewed.
How error docs stay up to date
Section titled “How error docs stay up to date”Error documentation and cost estimates are both delivered remotely through the same update system — no reinstallation needed. When a new error type is identified and documented, the updated messaging is pushed to all users automatically.
The next time the plugin starts (or within the regular refresh cycle), it checks for updates and applies any new error documentation. From that point on, what was previously a generic message becomes a specific, helpful one — with tailored fix steps and a “Read more” link.
No action is required from you. The plugin simply gets better at explaining what went wrong, without you doing anything.
How we find and fix new errors
Section titled “How we find and fix new errors”When something unexpected happens, modelBridge shows a safe, helpful message right away — you’re never left staring at raw technical output.
If anonymous error reporting is enabled (it’s on by default), the issue is automatically reported to the development team. No action is needed from you — the report contains only the error type and basic technical context, never your prompts, media, file paths, or any personal information.
The team reviews reported errors and publishes specific error messages and fix steps. These arrive the next time you open the plugin — what was a generic message becomes a clear explanation with recovery steps.
You can disable anonymous error reporting at any time in Settings > Privacy. The plugin works exactly the same either way — the only difference is whether the team learns about new issues automatically or has to wait for a manual report.
If an error persists
Section titled “If an error persists”If you keep running into the same error:
- Check the Error Reference for detailed information about specific error types, including causes and fix steps
- Review Common Issues for frequently encountered problems and their solutions
- Use the Send feedback option in the plugin’s API & Support tab to report the issue — include the model name and what you were trying to do