Frequently asked questions
Practical guidance for using NeonAPI.
Can I use a ChatGPT or Claude subscription as the API key?
No. Use the key created in the NeonAPI console. Subscription access and gateway API usage are separate.
Why is a model visible but my request is rejected?
Check the endpoint, exact model ID, key restrictions, group access and available quota. Model visibility does not imply support for every protocol or parameter.
Why is the charge different from a homepage comparison?
Homepage comparisons show reference rates under stated conditions. Actual charges can depend on group, input/output mix, context tiers, cache usage and task-specific units. Inspect the model details and usage record.
Does an error always mean no charge?
Not necessarily. An upstream may already have processed a request before the connection fails. Inspect the request or task record before retrying.
Can I paste my real key into these docs?
No key is needed to read the docs. Copy the placeholder examples into your trusted terminal or application environment and configure the key there.
Where do I find an unavailable feature?
Check Models & pricing and the console first. Only use capabilities published for your selected model; do not infer availability from another provider's documentation.