Automatic tracker of AI-token expenses (n8n)

I created an automation that tracks the usage of AI models (GPT-4, GPT-3.5, Claude) and shows where the AI budget is actually being spent. The system works as a separate "monitoring layer" and can connect to any n8n workflows.

How it works:
automatically collects data on AI model calls in the workflow;
counts prompt and completion tokens for each model;
groups expenses by GPT-4, GPT-3.5, Claude;
stores all statistics in Google Sheets in an understandable format.

Result for the business:
complete transparency of AI expenses;
clear understanding of which models and prompts are "eating" the budget;
no manual checks and surprises in expenses;
easier to control and scale AI systems.

Tools: n8n, n8n API, Google Sheets, OpenAI, Anthropic.