Error messages and troubleshooting
If a task stops with an error, read the message in the conversation and follow the suggested action. This page explains common error codes and steps you can try in WorkBuddy AI.
Understand an error message
An error message may include a title, a short explanation, an error code, and action buttons. The available actions depend on the problem and your app version. Use the message and actions shown in your app; a code alone may not identify the exact cause.
If your code is not listed below, follow Unlisted errors and feedback.
Common actions
Use an action only if it is available in your app. Button wording may vary by version or interface language.
| Action | What it does |
|---|---|
| Retry | Tries the step that failed again. |
| Switch model | Lets you choose another available model before trying again. |
| Check network | Opens the network check, when available. |
| New task | Starts a separate task when the current task or its context needs a fresh start. |
| Sign in again | Refreshes your sign-in session when the message asks you to sign in. |
| Configure model | Opens the settings for a custom model so you can complete its configuration. |
| Send feedback | Opens the feedback flow so you can report the problem. |
| View docs | Opens the help documentation. |
Network troubleshooting
For a connection failure, timeout, or other network-related message:
- Check whether other websites or online services work on the same computer.
- If you are using a work or school network, ask your IT team to check the connection and any required proxy settings.
- On a personal network, check the connection and, if possible, try another stable network. For a timeout, pause large downloads or uploads before trying again.
- If you use a proxy, check that its settings match your network requirements. Follow your organization's instructions for managed network settings.
- Use Check network if your app offers it. Once the connection is working, follow the task's message or use Retry if available.
If the banner says the task will resume automatically when the network recovers, you can wait for the connection to return and check whether the task continues before retrying manually. Follow the controls shown in your app if you want to stop waiting.
If the problem keeps happening, include the error message and the results of these checks in your feedback.
Common error codes
1001
What it may mean: The error is often related to the model service's state.
What to do: Use Retry to try again. If it still fails, wait a moment and try again later. If the problem continues, send feedback.
3002
What it means: The network connection failed. Your network environment may be involved.
What to do: Follow Network troubleshooting. If your message says the task will resume automatically, check whether it continues after the connection recovers.
3003
What it means: The network connection timed out. Your network environment may be involved.
What to do: Follow Network troubleshooting. If large uploads or downloads are running, pause them and try again on a stable connection. You can also try another model if one is available.
3007
What it may mean: The error is often related to the network environment.
What to do: Follow Network troubleshooting, then retry if your app offers that action. Send feedback if the problem continues.
3008
What it means: The model did not return a usable response or action instructions, so the task cannot continue for now.
What to do: Try again later, or switch to another available model. Send feedback if this happens repeatedly.
4100
What it means: Required connection settings for the selected custom model are incomplete, such as the API address or an API key required by the selected service. The request cannot be sent until the required configuration is complete.
What to do: On the desktop app, go to Avatar > System Settings > Models. Complete the fields required by your selected provider, save the configuration, and use Test Connection before trying again. API key requirements depend on the service; for example, the local Ollama preset does not require one. See Model Configuration.
Keep your API key private. Do not include it in feedback or screenshots.
6003
What it means: Requests are being rate limited.
What to do: Switch to another available model and try again, or try again later. Send feedback if the problem continues.
6004
What it means: Requests are being rate limited.
What to do: Switch to another available model and try again, or try again later. Send feedback if the problem continues.
11115
What it means: The input is too long.
What to do: Shorten the input, split it into smaller parts, or choose an available model that supports a larger context.
11133
What it may mean: The error is often related to the model service's state.
What to do: Start a new task, or switch to another available model and try again. Send feedback if the problem continues.
11134
What it may mean: The error is often related to the model service's state.
What to do: Switch to another available model and try again, or use Retry. Send feedback if the problem continues.
14003
What it may mean: The error is often related to the model service's state.
What to do: Switch to another available model and try again. If you cannot switch models, try again later. Send feedback if the problem continues.
Unlisted errors and feedback
If your code is not listed, or no code is shown, follow the error message and use the actions offered in your app. Use the network troubleshooting steps when the message describes a connection problem.
If you still need help, use the feedback action in the error message if available. On the desktop app, you can also go to Avatar > Help and Feedback, or Avatar > System Settings > About > Submit Feedback. See Help and Feedback for instructions and another way to contact us if in-app feedback is unavailable.
Include the following information:
- The error code, if shown, and the full error title or message.
- Your WorkBuddy AI version and operating system.
- The approximate time of the error and your time zone.
- The steps that led to the problem and what you have already tried.
- Whether you were using a work or personal network, if the problem was network-related.
- A screenshot with private information removed, if it helps explain the problem.
Do not include passwords, API keys, access tokens, private task content, or confidential files. If the feedback flow offers diagnostic information, review it before sharing and include only what is needed to investigate the problem.