Skip to main content
Your plan includes monthly automation runtime, AI, email, and web scraping allowances shared by every project in your organization. Usage resets at the start of each Coordinated Universal Time (UTC) calendar month.

​
Measure automation runtime

Automation runtime counts the elapsed time of each sandbox invocation. A trusted worker measures outside the sandbox in whole milliseconds, sums those measurements, and then displays the total in minutes. User code can’t change the timer, and Automate.ax doesn’t round each invocation to a full minute. Queue time, durable waits, scheduled delays, deployment planning, and time between invocations don’t count. Successful runs, failed user code, and timeouts count the runtime they consume. If the worker stops before it can record a measurement, that invocation is free. Free organizations stop starting new invocations after reaching the monthly allowance. An invocation already in progress can carry the total past the allowance. Paid organizations continue running while Automate.ax records usage beyond the included amount. The runs table groups contexts with the existing causal model and shows their combined recorded runtime. A three-second average is only a pricing estimate: 50 runtime minutes is about 1,000 such runs, while 2,500 minutes is about 50,000.

​
Count Platform AI usage

Platform AI measures the model cost of evaluate, generate, and generateImage calls that omit account and use the Automate.ax AI Gateway account. Automate.ax records the cost reported by the gateway after each request, so the final in-flight request can take usage slightly past the allowance. These costs remain separate from automation runtime. Calls that use your own provider account don’t consume Platform AI.

​
Count Platform email sends

Platform email counts sends from the built-in sendEmail action. Each recipient counts as one send. Sending email through integrations such as Gmail, Outlook, Resend, or Brevo doesn’t consume Platform email.

​
Count web scraping usage

Web scraping measures the browser milliseconds Cloudflare reports for each scrapeWeb attempt, including failed provider responses. Automate.ax rounds sub-millisecond values up to a whole millisecond and displays the total in minutes without rounding each scrape to a full minute. scrapeWeb stops starting new browser work after a finite monthly allowance is exhausted. A request already in progress can carry the total slightly past the allowance because its exact browser time is known only after Cloudflare responds. Cloudflare integration actions that use your connected account don’t consume the Automate.ax web scraping allowance.