Models, keys, and token budgets

Settings → Models & LLM access is one page covering three decisions: which model to use, whose key pays for it (a plan, or your own API key), and the monthly token ceiling you are willing to spend. Recent usage is summarized alongside them.

• The choice applies to both LLM nodes in your graphs and the AI graph builder.
• A bring-your-own-key setup keeps requests on your provider account.
• Hitting the ceiling stops LLM calls — raise it deliberately rather than as a reflex when generation starts failing.