> ## Documentation Index
> Fetch the complete documentation index at: https://docs.aicamp.so/llms.txt
> Use this file to discover all available pages before exploring further.

# Known Issues & Troubleshooting

> This page lists known issues you may occasionally experience while using AICamp, along with clear explanations and simple steps to resolve them. Most issues are temporary or related to AI model behavior and can be fixed quickly using the recommended actions below.
If you don’t find what you’re looking for, our support team is always here to help.

<Accordion title="1: Response was filtered due to content policy (400 error)">
  <Info>
    The response was filtered due to the prompt triggering content management policy. Please modify your prompt and retry. See our documentation for more information on content filtering policies.
  </Info>

  ### What does this mean?

  You may see this message when the AI does not return a response and instead shows an error related to Azure's content policy filtering.

  This does **not** mean your request is wrong or unsafe.

  In some cases, the AI model mistakenly flags a message because of certain words, phrasing, or context—even when the request is completely valid. This happens more often with **Azure-based AI models**, which apply stricter automated checks.

  ### Why does this happen?

  * The wording of the prompt triggers an automatic safety check
  * Long or complex instructions are misinterpreted
  * Azure AI models apply stricter filtering rules

  ### How to resolve this

  **Option 1: Switch the AI model**

  * Change the model from **Azure OpenAI** to **Direct OpenAI (GPT)**
  * Your chat history and data will remain unchanged

  **Option 2: Rephrase your prompt**

  * Use simpler language
  * Remove or reword sensitive terms
  * Break one long request into smaller steps

  **Option 3: Contact support**

  * Email us at [**support@aicamp.so**](mailto:support@aicamp.so)
  * Share what you were trying to do, and we’ll assist you

  ### Good to know

  * This is a known limitation, not a system failure
  * No data or conversations are lost
  * Switching the model usually resolves the issue instantly
</Accordion>