The baseline workload model assumes successful standard action steps at 1 task each. Zapier does not count triggers, Filters or Paths, failed/halted action steps, or several built-in tools as tasks. AI by Zapier can use 1x, 3x or 5x task rates depending on model tier; MCP, Lead Router and extended Code runtime can also use different rates.
Most non-AI module operations use 1 credit. Trigger and search modules normally use 1 credit per run; action modules can consume 1 credit per input bundle processed. Routers, filters and error-handler modules do not consume credits. Some AI features use dynamic rates, and Make Code uses credits based on execution time.
A workflow execution is one workflow run, regardless of how many nodes or steps it contains. n8n states that all listed plans include unlimited users and workflows. Failed executions do not count against the Cloud execution quota; paid-plan manual editor/chat executions and sub-workflow executions are also excluded from that quota.
Some provider links are affiliate links. AutomationCosts may earn a commission at no extra cost to you.
A note on price accuracy. Some providers publish a complete tier table, while others expose only entry prices or dynamic prices for a selected usage level. AutomationCosts stores exact prices only where the official source makes the tier explicit.