How to get the ChatGPT module reads a file

How To
,
1

I’m currently building a Make.com workflow where I retrieve tasks from ClickUp, download the attachments, and then send these documents to ChatGPT for copy improvements. I’m storing the attachments in a vector store, and they appear to be saving correctly. However, the ChatGPT module is unable to read these files—almost as if there’s a permission or access issue.

I tested the same prompt in ChatGPT’s Playground, and it works fine when referencing the stored documents. The problem arises only when I run the full workflow in Make.com. Has anyone experienced a similar issue or have any suggestions for resolving permission/access problems with ChatGPT reading attachments stored in a vector store on Make? or Any other better way to improve this workflow?

Thanks in advance!

This is my general workflow:

image
image1359×195 55.3 KB

Configuration of the ChatGPT module:

image
image526×658 45.6 KB

image
image499×925 58.6 KB

I put a timer to give time for the files to refresh but it still not working.

Error I am getting

image
image451×441 10.6 KB

2

Welcome to the Make community!

The File IDs field is a field that likely accepts an ARRAY.

A hint is the field name is plural with an “s”. That means it accepts multiple values and they all have to be within an array.

Even if you have a single item, you still need to map an array within that field. One quick way of doing that is to use the built-in function add with the special variable emptyarray.

{{ add(emptyarray; 46.id) }}

e.g.:

Hope this helps! Let me know if there are any further questions or issues.

— @samliew

P.S.: Investing some effort into the Make Academy will save you lots of time and frustration using Make.