[RESOURCE] Prompt Management in Airtable + Make.com

Sep 19 (edited) in 🤖 AI

If you use a lot of prompts within Make.com you know how annoying it is to update a large number of prompts at once such as when a new model comes out.

This is a simple resources that shows you how to easily manage your prompts within Airtable. See the first image for an example.

I've attached two Make.com scenarios:

1️⃣ This scenario uses webhooks to act as your own internal API. It will simply fetch all of the prompts from Airtable and return them back in JSON format. See the second image attachment.

I generated a UUID (for free via uuidgenerator.net) to act as an API key. This is passed in the URL and is checked before fetching the prompts. This is a good practice for any webhooks you have.

2️⃣ This scenario is an example of how to use the internal API. With one simple HTTP GET request you get back a JSON object like this:

{

"promptName": {

"model": "gpt-4o",

"max_tokens": 0,

"system": "The system prompt will be here"

"anotherPromptName": {

// etc

}

You can then easily access it within an OpenAI module (see the fourth image attachment).

🔥 I've attached the Make.com scenarios below so you can download them. The Airtable setup is simple enough to replicate manually.

💡 Let me know if you have other ideas on how to improve this system!