Track AI Token Usage Metrics for OpenAI & Gemini in n8n

This n8n workflow is designed to track and monitor AI token consumption in workflows utilizing OpenAI and Gemini models. By providing an execution_id and an array of AI node names (e.g., ['openai', 'gemini']), it fetches the full execution data, scans the specified nodes for token usage details, aggregates prompt, completion, and total tokens, and compiles a list of models used. The output includes comprehensive totals and a detailed breakdown per model, making it easy to analyze resource usage post-execution.

To use it effectively, configure the workflow to run at the end of your main AI workflow with 'Wait for completion' disabled to ensure full execution data is captured. This prevents blocking and guarantees accurate metrics. It's ideal for developers and teams building AI-driven automations who need reliable insights without manual logging.

Benefits include precise cost forecasting (since tokens directly impact API bills), prompt optimization by identifying high-usage patterns, and usage auditing for compliance or performance tuning. Use cases span AI content generation in publishing, media automation for scriptwriting or image prompting, dev tools for testing model efficiency, and enterprise AI pipelines where token budgets are critical. Save hours of manual data extraction and enable data-driven decisions to reduce expenses by 20-50% through better model and prompt management.