Common errors and their solutions
1. Context Too Long
The AI has reached the context length limit. Please create a new conversation or try again with a smaller prompt.
Explanation: The AI models used by Mark AI have a technical limitation regarding the amount of information they can process simultaneously, known as the "context."
💡 Even when this limit is not completely reached, the model's performance decreases as the context gets longer.
ℹ️ The context includes all messages exchanged in the conversation as well as imported documents.
Solution:
Reduce the number of documents imported in your last message.
Start a new conversation if the problem stems from the number of messages already exchanged.
2. AI Overloaded
The AI is currently overloaded, meaning it's experiencing too high a volume of requests. Please try again in a few minutes.
Explanation: This error comes from the AI providers used by Mark (Amazon Web Services, Microsoft Azure, Google Cloud Platform). It appears when the provider doesn't have the capacity to handle all the demand. In case of an extended outage or overload, Mark AI can temporarily switch assistants to other AI providers or models.
Solution:
Wait a few minutes before trying your request again.
If the problem persists beyond 15 minutes, contact Mark AI customer support to request a temporary switch to another AI provider; we'll try to respond as quickly as possible.
3. No Content Generated
The AI did not generate any content. Please try again by modifying your instruction.
Explanation: This error occurs when the AI model generates an empty message or stops unexpectedly. This can be due to an internal model error or a misunderstanding of your request.
Solution:
Restart the generation.
If the error persists, slightly rephrase your request to make it clearer or more precise.
Check that your request doesn't contain elements that could be misinterpreted by the AI.
4. Quota Limit Reached (Rate Limit)
You have reached the generation quota; you can generate again in [number of hours] hours and [number of minutes] minutes.
Explanation: Mark AI imposes a limit on the number of messages you can generate per 4-hour period. This limit varies according to the type of license you have. Discovery licenses have a significantly lower limit than unlimited licenses.
⚠️ If you reach this limit with an unlimited license, it probably indicates abnormally intensive content generation.
Solution:
Wait for the indicated time before you can generate content again.
If you are using a discovery license, consider upgrading to an unlimited license to benefit from a higher quota.
5. Image Limit Reached
You have reached the limit of [number of images] images per conversation.
Explanation: Each AI model used by Mark AI has technical limitations regarding the number of images it can process in a single conversation. These limits vary from one model to another.
💡 Even before reaching this limit, the model may lose reliability, fail to take certain information into account, or misinterpret your instructions if it contains many images and documents.
Solution:
Reduce the number of images in the message you want to send.
Start a new conversation including only the images essential to your request.
If you absolutely must use multiple images, distribute them across different conversations.
6. Document Limit Reached
You have reached the limit of [number of documents] documents per conversation.
Explanation: As with images, AI models have technical limitations regarding the number of documents they can process in a conversation.
💡 Even before reaching this limit, the model may lose reliability, fail to take certain information into account, or misinterpret your instructions if it contains many images and documents.
Solution:
Reduce the number of documents in your current message.
Start a new conversation including only the essential documents.
For large documents, consider importing only the relevant sections rather than the entire document.