Ask AI Errors Reference
These error codes correspond to issues encountered when using Ask AI. Each code includes a short description and possible resolutions.
General API errors
These error codes represent issues at the HTTP or service level, independent of the AI stream. They cover authentication, routing, rate limiting, and server errors.
AI-201 - Bad input
The request included data that could not be used to find a resource, or the data was malformed.
Solution: Make sure you are setting the correct configuration options for DocSearch or SiteSearch, and that you are passing the correct request body in your API call.
AI-202 - Unauthorized
There was an issue validating the authentication token for the request.
Solution: Ensure that you are including a valid authentication token in your request headers and that it has not expired.
AI-203 - Forbidden
Access to the requested resource is unavailable or forbidden for this requester.
Solution: Verify that your domain is whitelisted, that you are using the correct assistantId, and that your account has the necessary permissions for this resource.
AI-204 - HTTP error
There was an issue with HTTP routing or endpoint configuration. This may indicate a problem with the target URL, proxy setup, or request path.
Solution: Double-check that the API endpoint is correct and that your network or proxy configuration allows outbound connections.
AI-205 - Too many attempts
Too many requests were made within the designated time window. Your requests are being rate-limited.
Solution: Wait for the rate limit window to reset before sending additional requests.
AI-206 - Unknown
An unknown error occurred.
Solution: Retry the request. If the issue persists, contact support with the full request and response details.
AI-207 - Internal error
An internal error occurred within the Ask AI service.
Solution: Retry your request later. If the problem continues, contact support with the error ID or timestamp.
Ask AI chat errors
These error codes represent issues with the AI stream or upstream LLM providers.
AI-208 - Provider API call error
There was an issue communicating with the upstream provider's API. This can occur when the provider returns an unexpected response or a timeout.
Solution: Check the provider's API status page(s) to see if they are experiencing any ongoing issues.
AI-209 - Provider retry error
There were too many failed attempts at communicating with the upstream provider. This could be caused by issues on the provider's end, an invalid API key, or a network issue causing repeated retries.
Solution: Check the provider's API status page(s) to see if they are experiencing any ongoing issues.
AI-210 - No tool available
The AI runtime could not call the requested tool. This usually means the tool name referenced in your request does not exist or was not registered in the SDK configuration.
Solution: Be sure not to ask for any specific tool calls within your custom prompt. Ask AI has specific tools it uses to give the best results.
AI-211 - Stream error
A general error occurred while streaming data from the upstream provider. This could include connection interruptions, malformed data, or provider-side streaming issues.
AI-212 - Invalid API key
The upstream provider reported that there was an issue with the supplied API key.
Solution: Make sure to double-check that you are using the correct API key for your assistant, and that it has the correct permissions.
AI-213 - Billing error
The upstream provider could not process the request due to a billing or account balance issue. This includes insufficient balance, credit limits, or account payment problems.
Solution: Log in to your provider's billing dashboard to check your account balance, payment methods, and billing status. For Anthropic users, visit Plans & Billing to upgrade or purchase credits.
AI-214 - No existing model
The selected model does not exist for the chosen provider.
Solution: Make sure the model exists for the selected provider. A list of supported models and providers can be found in our docs.
AI-215 - No existing provider
The selected provider does not exist.
Solution: Make sure the provider is supported. A list of available providers and their compatible models can be found in our docs.
AI-216 - Could not load API Key
This error occurs when the SDK could not load the provider's API key successfully.
Solution: Ensure the API key for the chosen LLM provider is correct and has all of the correct permissions added to it.
AI-217 - Thread Depth Exceeded
The conversation has reached its maximum depth limit and can no longer accept follow-up questions.
Solution: Start a new conversation to continue. The conversation depth limit is configured per assistant to maintain response accuracy and prevent conversation drift.
AI-218 - Invalid custom provider
The custom provider configuration is invalid or missing required fields. This occurs when using the custom-openai provider without proper configuration.
Solution: Ensure your custom provider configuration includes a valid
baseUrland any required custom headers. Check the custom provider documentation for configuration requirements.
AI-219 - Custom provider retry failed
A request to a custom provider failed after exhausting all retry attempts. This can occur due to network issues, timeouts, or the custom provider being unavailable.
Solution: Verify that your custom provider endpoint is accessible and responding correctly. Check network connectivity and the custom provider's status. Review timeout and retry settings in your configuration.
AI-220 - Quota exceeded
The provider's rate limit or quota has been exceeded. This can happen when you've made too many requests in a short time period or exceeded your plan's usage limits.
Solution: Check your provider's dashboard to review your current usage and rate limits. For Google Gemini, visit the quota monitoring page. For OpenAI, check your usage dashboard. You may need to wait for the quota to reset or upgrade your plan.
AI-221 - Model not found
The specified model does not exist or your API key does not have access to it. This can occur when using a model that requires special access permissions or has been deprecated.
Solution: Verify that the model name is correct and that your account has access to it. Check your provider's dashboard for available models and access permissions. If you recently gained access to a new model, it may take a few minutes to propagate.
AI-222 - Region mismatch
The request was made to an incorrect regional endpoint or the resource is restricted by geography requirements. This typically occurs with OpenAI when using regional data residency settings.
Solution: When creating a new assistant with OpenAI in the Dashboard, select the appropriate data residency option that matches your requirements (none, Europe, or United States). If you're experiencing this error with an existing assistant, verify the data residency setting in your assistant configuration and ensure it matches your account's regional requirements.
AI-223 - Verification required
Your organization must be verified to access certain provider features. This is commonly required for OpenAI's advanced features like reasoning summaries.
Solution: Visit your provider's organization settings to complete the verification process. For OpenAI, go to organization settings and click "Verify Organization". Note that after verification, it may take up to 15 minutes for access to propagate.
AI-224 - Context length exceeded
The request exceeded the model's maximum context length. This happens when the combined size of your messages and requested completion exceeds what the model can handle.
Solution: Reduce the size of your conversation history or input. You can either start a new conversation, remove older messages, or shorten your prompts. Check your model's token limits in the models documentation.