Can I use my own AI API key on n8n Cloud?

Yes. Bringing your own AI provider key (Anthropic, OpenAI, etc.) for AI nodes in your workflows has always been available on every n8n plan and edition — Cloud Starter, Pro, Enterprise, and self-hosted.

Using your own key on an AI node

On any node with an AI credential (OpenAI, Anthropic, etc.), open the credential field and select "Create new credential", then add your provider API key as usual to set up your credential.
This works the same way on every plan. It has no plan-tier requirement and is not related to or dependent on Gateway credits, AI Assistant credits, or instance-level MCP.

Gateway credits (the optional alternative)

Gateway credits let you run supported AI models and tool services without creating a provider account or API key — n8n bills usage from a prepaid balance instead.

Available on n8n Cloud Starter and Pro only.
Available from n8n 2.36.0 onward.
Choosing Gateway credits is always optional — you can switch any supported node back to your own credential at any time.

To use Gateway credits, you need to make sure the feature is enabled for your instance. You can do this by going to Admin Panel > Manage > Enable Gateway credits.

Also, make sure that you have Gateway credits available to use them. You can verify this by going to Admin Panel > Manage > Gateway credits. If needed, you can top up your Gateway credits and enable auto-top-up.

Not the same as AI Assistant credits or instance-level MCP

Two other n8n features are easy to mix up with the above:

AI Assistant credits are what power your AI Assistant sessions. See Can I add or reset my AI credits/tokens? for the reset policy for that feature.
Instance-level MCP lets external AI clients (Claude Desktop, Claude Code, ChatGPT, Cursor, etc.) connect to your n8n instance to search, execute, and build your workflows, among other things. It's available on every plan and edition. See this doc for setup.