FAQ
Many common issues can be fixed in seconds with the answers below. Keep it bookmarked for quick reference.
Running Agents
Cause — This usually happens when the prompt is empty or the model times out.
Fix
- Make sure your internet connection is stable.
- Keep prompts under 2,000 characters for smoother runs.
- Reload the page and retry.
If issues persist, check if your daily message limit is reached (see Plan Limits).
Cause — Some third-party APIs are slower than Weaver’s default wait time.
Fix
- Increase the Maximum Wait Time setting in the APICall component.
- For very long-running jobs, switch the workflow to an Async → Await pattern so execution can resume once the API responds.
Debugging Agents
Cause — Fix with AI only appears when a component is selected and logs show an error.
Fix
- Select the failing component with a red border.
- Or multi-select with Cmd / Ctrl.
- Run again with Debug Mode enabled.
See Debugging with Weaver for full troubleshooting steps.
Cause — Debug Mode was not enabled.
Fix
- Toggle Debug Mode in the top toolbar before pressing Run.
- Once active, logs will capture inputs, outputs, and errors for every component.
Read more in Agent Logs.
Cost and Plan Limits
Free plans allow up to 5 messages per day in Weaver. After that, you’ll need to wait until the next day or upgrade your plan to continue building and testing agents.
Tip — If you hit limits often, move to a paid plan so you can iterate without interruptions.
See Billing Management for upgrade options.
Cause — Some skills or prompts may be generating unnecessarily long outputs.
Fix
- Go to Analytics → Model Usage to identify which skill is responsible.
- Optimise prompts to keep responses scoped.
- Choose lighter models (e.g., GPT-4o mini) when high accuracy is not required.
- Add constraints to outputs (e.g., “Return JSON only” or “Limit response to 200 words”).
See Best Practices for more optimisation tips.
Using Features
Drag and drop a PNG, JPG, or SVG directly into the chat. Weaver will convert the file to base64 and pass it to the appropriate image-processing component automatically.
Hold Cmd / Ctrl while clicking each component, or drag a selection box starting from an empty space on the Canvas.
Getting Help
If none of the above fixes solve your issue:
- Export your agent and test it locally with the SmythOS runtime.
- Share logs and repro steps in the #support channel on Discord.
- The support team and community can help you resolve complex issues faster.