Enter agent runs, app actions, and MCP tool calls per run, and Model Ruler estimates the iPaaS platform layer — modeling Zapier MCP tool-call accounting and module fan-out — separately from LLM inference, which belongs in the agent loop calculator. Deterministic and source-dated, not a quote.
Estimates the automation-platform layer for agent workflows by modeling app-action fan-out and MCP tool-call accounting separately from LLM token cost.
Formula: zapier_tasks = agent_runs * (app_actions + 2 * mcp_tool_calls); make_credits = agent_runs * make_modules; n8n_executions = agent_runs
Audit coverage: Automation platform billing · Agent-loop fan-out
What does this calculator compute?
It models the automation-platform layer of an agent workflow — app-action and MCP tool-call fan-out across Zapier, Make, and n8n public-plan floors — and intentionally excludes LLM token spend, which the agent-loop calculator covers.
What should I check before committing to a workload shape?
Per-step model calls plus any automation/iPaaS module actions both accumulate. Separate the LLM cost from the platform per-operation cost — a workflow can look cheap on tokens while the per-module automation charges become the larger line at volume.
Is this a quote or a benchmark?
Neither. It is a computed estimate. Real-world cost varies by rate-limit shape, caching, token-count precision, and provider-specific discounts not reflected in the public pricing table.
Zapier, Make, and n8n public pricing floors reviewed from official source pages. This estimate normalizes tasks, credits/modules, and executions; confirm current provider pricing before purchase or production commitments. View methodology.