AI Assistant isn't aware of trial plan limits (recommends creating an n8n API key, which isn't available on trial)

Summary

The AI Assistant did a great job building a workflow from a natural language prompt on my first try with n8n Cloud (previously self-hosted). One issue: it instructed me to create an n8n API key as part of the setup, but API keys aren’t available on the free trial. The assistant doesn’t seem to have my plan or entitlements in its context, so it recommends steps I can’t actually complete.

Steps to reproduce

  1. Sign up for an n8n Cloud free trial.
  2. Open the AI Assistant and ask it to build a workflow that requires the n8n API (e.g., one that reads or writes workflows via the n8n API node, or otherwise needs an API key for the instance itself).
  3. The assistant tells you to go create an n8n API key.
  4. Go to Settings > n8n API. The API key section isn’t available, since it’s a paid-plan feature.

Expected behavior

The assistant should know the current plan’s entitlements and either:

  • Flag upfront that this step needs a paid plan, so the user isn’t sent on a dead-end, or
  • (further out) create the API key itself, if it already has enough permission/context to do so.

Actual behavior

The assistant proceeds as if the API key is available, with no mention of the plan restriction. The user only discovers the block when they try to follow the instructions.

Impact

Minor but avoidable friction on a first-time trial experience. I’d guess this doubles as a soft paywall/conversion nudge, which is a reasonable business call, but making it explicit (“this step needs a paid plan”) would be a better experience than a silent dead end.

Suggested fix (two tiers)

  1. Short term: surface plan/entitlement awareness in the assistant’s context so it can flag gated steps before sending the user to a dead end.
  2. Longer term: let the assistant create the API key on its own when it has the necessary permissions, removing the manual step entirely.

Environment

  • n8n Cloud, free trial
  • Previously used self-hosted n8n (for context, in case behavior differs from self-hosted assumptions)