n8n Sub-Workflow: AI Model Token Usage and Cost Tracker (Jina AI)

This n8n sub-workflow is designed to track and log AI model token usage and associated costs, specifically for Jina AI models, into a Google Sheet. It acts as a backend service that can be called by other workflows whenever an AI model interaction occurs.

Description

This workflow simplifies the process of monitoring AI model consumption by receiving token usage data, calculating the cost based on a predefined rate, and then appending this information to a designated Google Sheet. This allows for easy tracking and analysis of AI service expenditures.

What it does

1. Receives Data: It starts by listening for execution requests from another workflow, expecting input data containing details about AI model usage (e.g., model name, input tokens, output tokens, timestamp).
2. Calculates Cost: A Code node processes the received token counts (input and output) and, based on hardcoded rates for Jina AI models, calculates the total cost of the AI interaction.
3. Prepares Data for Google Sheets: An "Edit Fields (Set)" node structures the data into a format suitable for appending to a Google Sheet, including the model name, input tokens, output tokens, total tokens, and calculated cost.
4. Logs to Google Sheets: Finally, it appends the prepared token usage and cost data as a new row to a specified Google Sheet.

Prerequisites/Requirements

• n8n Instance: A running n8n instance to host and execute the workflow.
• Google Account: A Google account with access to Google Sheets.
• Google Sheets Credential: An n8n Google Sheets credential configured to allow writing to your desired spreadsheet.
• Jina AI Model Usage: This workflow is specifically designed to track usage from Jina AI models, with hardcoded token pricing. If you use other models or providers, you will need to adjust the pricing logic in the "Code" node.

Setup/Usage

1. Import the Workflow: Import this JSON definition into your n8n instance as a new workflow.
2. Configure Google Sheets Node:
   • Select your Google Sheets credential.
   • Specify the Spreadsheet ID of the Google Sheet where you want to log the data.
   • Specify the Sheet Name (e.g., "Sheet1").
   • Ensure the first row of your Google Sheet contains headers matching the keys output by the "Edit Fields (Set)" node (e.g., model, inputTokens, outputTokens, totalTokens, cost).
3. Review and Adjust Code Node (Optional):
   • The "Code" node contains the pricing logic for Jina AI models. If Jina AI changes its pricing, or if you wish to track other models, you will need to update the modelPrices object within this node.
4. Activate the Workflow: Set the workflow to "Active" to enable it to receive calls.
5. Call from Another Workflow: This workflow is designed to be called as a sub-workflow using an "Execute Workflow" node from another n8n workflow. The calling workflow should pass an item with the following structure:

   {
     "model": "jina-embeddings-v2-base-en",
     "inputTokens": 100,
     "outputTokens": 0,
     "timestamp": "2023-10-27T10:00:00Z"
   }
   

   (The timestamp is optional as the workflow adds its own, but can be overridden.)

This setup allows you to centralize your AI token usage tracking across various n8n workflows that interact with Jina AI models.