Chat error messages

When a chat request can't be answered, meinGPT shows a specific error message. Most of these errors are temporary and originate with the AI provider (e.g. OpenAI, Anthropic, Google) — not with your account or your usage. This page explains every known error message: what it means, why it happens and what you can do.

Every chat error message includes a "Why am I seeing this error?" link that takes you straight to the matching section on this page.

Rate limit reached [#rate-limit-exceeded]

What does it mean? The AI provider has reached a request (rate) limit for the selected model. These limits apply per time window and are set by the provider to share its capacity evenly.

Why does it happen? During peak times or for very popular models, the provider's capacity is briefly exhausted. This affects everyone using that model, not just you.

What can you do?

• Use Switch model to pick a different model — this usually works right away.
• Or wait a moment and try again via Retry.

Server issues at the AI provider [#provider-server-error]

What does it mean? The AI provider returned a server-side error (e.g. an internal failure or a timeout). The request couldn't be completed as a result.

Why does it happen? These outages are entirely on the provider's side and are usually short-lived.

What can you do?

• Try again via Retry.
• If the problem persists, use Switch model to move to a different provider/model.

Chat too large for the model [#context-length-exceeded]

What does it mean? The chat has grown longer than the current model can process at once (its "context window" is exceeded).

Why does it happen? Every model can only consider a limited amount of text at a time. Long chats, large documents or extensive attachments fill up this window.

What can you do?

• Use Switch model to choose a model with a larger context window.
• Or start a New chat and include only the information you actually need.
• Keep an eye on the context window indicator in the chat input — it shows how much of the window the current chat already uses, so you can switch or start fresh before you hit the limit.

Request blocked by the content filter [#content-filter]

What does it mean? The AI provider's safety/content filter blocked the request or response.

Why does it happen? Providers screen inputs and outputs for certain categories (e.g. violence, self-harm). Sometimes these filters trigger too strictly, even for harmless requests.

What can you do?

• Rephrase your request via Edit message.
• Or try a different model via Switch model — filters differ from provider to provider.
• Want to know what providers screen for — or think the block was justified? See their usage policies: OpenAI↗, Anthropic↗, Google↗, Mistral↗, and Perplexity↗.

Connection interrupted [#stream-interrupted]

What does it mean? The connection dropped while the response was being generated, before it fully arrived.

Why does it happen? This can occur with especially long responses or an unstable network connection (e.g. a timeout between browser and server).

What can you do?

• Try again via Retry.
• If it keeps dropping, a shorter prompt or a different model helps.

Maximum number of agent steps reached [#agent-max-iterations]

What does it mean? The AI agent reached the maximum allowed number of tool calls (steps) without completing the task. This is a built-in safety limit on steps — it has nothing to do with your credit balance, so it can happen even when you have plenty of credits left.

Why does it happen? Very complex or vaguely worded tasks can make the agent take too many intermediate steps.

What can you do?

• Break the task into smaller sub-steps.
• Make the request more specific and try again via Retry.

Model doesn't support tools [#unsupported-tool-use-for-model]

What does it mean? The selected model can't use tools (e.g. web search or document generation), but the request needed one.

Why does it happen? Not every model supports tool use. If a chat relies on tools, it only works with models that offer it.

What can you do?

• Use Switch model to pick a model that supports tools.

File type not supported [#unsupported-file-type]

What does it mean? The selected model can't process the type of file you attached.

Why does it happen? Models support different file formats, and the attached format isn't one this model accepts.

What can you do?

• Convert the file to a supported format (e.g. PDF) and add it again.
• Or use Switch model to pick a model that supports the format.

PDF too long [#max-pdf-pages-exceeded]

What does it mean? The attached PDF has more pages than the model can process.

Why does it happen? Models have a limit on the number of PDF pages per request.

What can you do?

• Use a shorter document or split the PDF into several parts.
• If needed, use Switch model to choose a model with a higher page limit.

Image too large [#image-size-exceeded-mb]

What does it mean? The attached image exceeds the maximum allowed file size (in MB).

Why does it happen? Models only accept images up to a certain file size. Some providers count the combined size of all attached images, so several small images together can also exceed the limit.

What can you do?

• Shrink the image or save it in a more compressed format.
• If you attached several images, remove some or split them across multiple messages.

Image resolution too high [#image-size-exceeded-pixels]

What does it mean? The attached image exceeds the maximum allowed resolution (in pixels).

Why does it happen? Models only accept images up to a certain pixel resolution.

What can you do?

• Reduce the image resolution and add it again.

Too many images [#image-count-exceeded]

What does it mean? The request contains more images than the model allows.

Why does it happen? Each model accepts only a limited number of images per request.

What can you do?

• Remove some images and try again.
• Or split the images across several messages.

Image too large for multi-image requests [#image-dimensions-exceeded]

What does it mean? Several images are attached and one of them exceeds the model's per-image pixel limit.

Why does it happen? For requests with multiple images, providers cap the resolution allowed per individual image.

What can you do?

• Reduce the resolution of the affected image and try again.
• Or send fewer images at once.

Document generation failed [#document-generation-error]

What does it mean? Converting the content into the target format (e.g. Word or PDF) failed.

Why does it happen? Certain content or formatting can't be converted reliably.

What can you do?

• Try again after a short while.
• Adjust the document slightly (e.g. remove unusual formatting) and regenerate it.

Premium model needs credits [#premium-model-credits-required]

What does it mean? The selected model is a premium model that requires purchased credits, and your workspace doesn't have any yet.

Why does it happen? Premium models are billed via credits. Without purchased credits, they can't be used.

What can you do?

• As a workspace admin, add credits via Open billing settings.
• If you're not an admin, ask a workspace admin to add credits.
• Or use Switch model to pick a model that doesn't require credits.

Workspace out of credits [#credits-exhausted]

What does it mean? Your workspace has used up all its credits.

Why does it happen? Credit-based models draw from a shared workspace balance. Once it reaches zero, further requests are blocked until more credits are added.

What can you do?

• As a workspace admin, top up via Open billing settings.
• If you're not an admin, ask a workspace admin to add credits.
• Or use Switch model to continue with a model that doesn't require credits.

Workspace usage limit reached [#member-text-tier-limit-exceeded]

What does it mean? You've reached your text usage limit for this workspace in the current billing period.

Why does it happen? Workspaces can set a usage tier per member. Once its limit is used up, further requests are temporarily blocked.

What can you do?

• Use Request upgrade to ask for a higher usage tier — a workspace admin can assign it.

Unknown error [#unknown-error]

What does it mean? An error occurred that can't be assigned to any known category.

Why does it happen? The cause is usually temporary (e.g. a brief outage). Technical details, if available, are shown directly in the error message.

What can you do?

• Try again via Retry, or start over with New chat.
• If the error keeps happening, contact support with the details shown.