Overview
Most problems are one of a handful of things. Find your message below. If a provider has its own guide, it lists more specific fixes.
Common errors
| What you see | What to do |
|---|---|
| “Your key was rejected” (401 or 403) | The key is wrong, revoked, or from a different place than the address (a different region, platform or organisation). Paste it again with no spaces, or create a new one. |
| “Model or endpoint not found” (404) | The model name is not available to you, or the address is wrong. Pick another model from the chip, and check the Endpoint under the provider. |
| A billing or balance message (402, “quota”, “credit”) | The account has no credit. Add some in the provider’s billing page. Creating a key is usually free; using it is not. |
| Rate limited (429) | You hit the provider’s limit, often on a free tier. Wait a minute, pick a lighter model, or upgrade. |
| “This chat is longer than the model can read” | Loom already tries to fold older turns into notes. If you still see this, start a new chat or choose a model with a bigger window. See Chat notes in the chat menu. |
| “The model sent back an empty reply” | Try again, or choose another model. Reasoning models can use their whole allowance thinking. |
| Can’t reach a server on your network | Same Wi-Fi? Right address and port? Firewall allowing the port? See the Ollama or LM Studio guide. |
| Tools or search don’t run | Not every model supports tools. Try a model that does, and check search or the connector is switched on in the + menu. Incognito chats never use tools. |
| A photo or PDF isn’t read | Not every model or host reads them. Loom tells the model when a file could not be sent. Try another model. |
| Replies stop short | The reply hit the length limit. Say “continue”, or choose a model with a larger output limit. |