When an article can’t be generated, SynaPost shows one short message saying what went wrong and what to do. You’ll see it on the Content Pipeline (hover over the icon on a failed topic), in the generation status next to the button you clicked, and in the Daily Automation log.
Here’s what each message means and how to fix it.
“The AI provider key is missing or was rejected.”
The key isn’t saved, or the provider refused it. Check it under Settings → AI & Images. Paste the key again, click Test Connection (or Test and connect for OpenRouter), and save. SynaPost stops right away on a rejected key, so it won’t keep retrying other models with the same bad key.
“The AI provider account is out of credits or has reached its spending limit, so this request cannot run.”
Your provider account has no credit left, or a spending limit you set has been reached. Add credits or raise the limit with the provider. Or switch to a different model under Settings → AI & Images. For OpenRouter, paid models are billed against your OpenRouter credits, and free models don’t need credits.
“Today’s free usage limit for this AI provider has been reached.”
A free provider tier has run out for today. Try again tomorrow, or connect a different provider under Settings → AI & Images. See Connecting OpenRouter for how free limits work.
“The AI provider is busy right now because of its usage limits.”
The provider is limiting requests for a short time, usually per minute. Wait a few minutes and try again. You don’t need to change anything. Automatic retries, if enabled, will try again on their own.
“This AI model is only available with SynaPost Pro.”
The model or provider you picked needs an active SynaPost Pro license. Pick a free model under Settings → AI & Images, or activate your license under Pro License.
“SynaPost couldn’t reach the AI provider.”
Your server couldn’t connect to the provider. Check that the server has internet access and that the provider isn’t blocked by a firewall. If you use Ollama, check that Ollama is running and reachable from the server. Then try again.
“The AI model returned no usable text this time.”
The model answered, but without any article text. Some reasoning-heavy models use their whole output budget on internal thinking and return nothing visible. Trying again usually works. If it keeps happening, pick a different model under Change model on the provider’s card.
“The AI model is temporarily unavailable right now.”
The model is overloaded, down, or restricted at the moment. SynaPost has already tried other models from the same provider. Try again later, or choose a different model under Settings → AI & Images.
“Content couldn’t be generated right now.”
Something unexpected went wrong that doesn’t fit the messages above. Try again in a few minutes. If it keeps happening, the error is also recorded in the generation log, so you can check the exact provider response there.
Automatic retries
If the automatic retry option is turned on under Settings → AI & Images, a failed topic is retried after a short delay, up to the number of attempts you set. The Pipeline shows when the next attempt is scheduled. Once the attempts run out, the topic stays Failed until you regenerate it.