How can I upload an entire Google Sheet as a file to OpenAI via n8n?

Below is some guidance on working around this limitation:

What’s Currently Possible

The OpenAI API’s file upload mechanism for batch processing only accepts files in a specific format (a .jsonl file with a max size of 200 MB). In n8n there isn’t a built-in node that directly exports your Google Sheet as a file in that format and sends it to OpenAI in a single API call. The typical approach is to process each row separately, but as you mentioned, this isn’t scalable for your use case.

Workaround Options

1. Export and Transform
   You can build a custom n8n workflow that:

   • Uses the Google Sheets node to retrieve all rows from your sheet.
   • Transforms the data (using a Function or Code node) into the required .jsonl format.
   • Saves the result as a file (for example, using the Write Binary File node).
   • Uses an HTTP Request node to send this file to the OpenAI Batch API.

2. Custom Implementation
   If you’d like the chat-like experience (upload a file with instructions and receive a modified file back), you might consider:

   • Exporting your Google Sheet data as JSONL (as above).
   • Calling your own API (or a Custom Workflow Tool in n8n) that wraps the OpenAI API call.
   • Processing the received output, then updating your Google Sheet accordingly.

Future Considerations

At present there is no direct, out-of-the-box feature in n8n to send an entire Google Sheet as a file to OpenAI. Adopting the export-transform-upload workflow is the most feasible option. While n8n is continuously evolving, there have been no announcements about a feature that directly uploads the entire sheet file to OpenAI as you interact with ChatGPT. However, feedback on use cases like yours is invaluable and may influence future enhancements.

Example Transformation (Pseudo-code)

Below is a brief example of how you might transform your sheet data to .jsonl format using a Function node:

JSCopied
// Assume `items` is an array of rows from Google Sheets, where each item is an object.
const jsonlLines = items.map(item => JSON.stringify(item)).join('\n');

return [{ json: { content: jsonlLines } }];

You can then pass the resulting text to a node that saves it as a file, and finally send it using an HTTP Request node with multipart form-data to the OpenAI Batch API.

---

By following these steps, you can simulate a “file upload” process within your n8n workflow even though a direct method isn’t currently supported.