Track AI Agent Token Usage & Estimate Costs in Google Sheets

This n8n workflow monitors token consumption from AI Agent executions, extracting data from workflow metadata since agents don't fully expose tool call details. It works seamlessly with OpenAI, Google Gemini, and Anthropic models, with minor tweaks for others. After the AI Agent runs, it triggers a sub-workflow to calculate precise token counts, then appends the data to a Google Sheet.

The Google Sheet includes built-in formulas to estimate costs based on provider pricing tiers, providing real-time visibility into expenses. This setup ensures comprehensive tracking even in complex agent interactions involving multiple tool calls.

Benefits include cost optimization, usage analytics for AI projects, and prevention of budget overruns. Ideal use cases: AI development workflows, production agent monitoring, SaaS apps with LLM integrations, and teams managing multiple LLM providers. Requirements: LLM API keys, Google Sheets/Drive credentials, and n8n instance API key.