We built AI Cost Guardrails-CNXT because AI usage in WordPress is moving beyond one-click actions into background jobs, Cron tasks, automated workflows, and plugin-driven processes. That creates a new operational risk: nothing has to be "hacked" for AI costs to run away. A retry loop, repeated request, oversized prompt, unexpected model change, or broken background job can consume a large share of the monthly budget before anyone notices. AI Cost Guardrails-CNXT adds a safety layer around calls made through the native WordPress AI Client. The Free tier provides monthly request and token boundaries, while Pro adds earlier controls for sudden token growth, repeated requests, retry storms, caller-specific budgets, model spend, per-request size, and execution context. We also designed it around data minimization: raw prompts and AI responses are never persisted. Our goal is simple: if WordPress is going to use more AI, it should also have a practical way to contain abnormal AI consumption. We would really value feedback from WordPress developers, plugin teams, agencies, and site operators. NOTES - Current scope: native WordPress AI Client lifecycle. Direct provider SDK/HTTP calls are outside scope. - Suggested first gallery image: English hero image with the clearest "AI cost guardrails" message.

AI Cost Guardrails-CNXT

Stop runaway AI costs in WordPress before they spread.

Stop runaway AI costs in WordPress before they spread.