I didn’t find any similar post so I will just create this new one. I am finding that the documentation is really outdated, which just makes it difficult for me (or the LLM agent) to work effectively with Flows.
Look at the beginner tutorial “Create a list-based loop with the For block”. OK, this sounds great. What should I do first? here’s the first five steps.
To create a new flow, do the following:
-
Choose an existing workspace or create a new one.
-
From the Postman sidebar, select
Flows. -
Select
Create a new Flow > Create new Flow. -
Enter a name. For example, enter Create a list-based loop.
-
Click the Return or Enter key.
Ok, let’s see which of these steps are still valid.
Step 1, I choose my workspace. Fine.
Step 2, I select Flows. Cool, works for me because I have enabled flows (too bad for those suckers who might have switched it off.)
Step 3, it says select “+ Create a new Flow”. Ok, is it just me but there is nothing that says “+ Create a new Flow” anywhere. There are three ways to create a new flow, but none of them say that. And when I create a flow, there’s not just new flow, it’s either “Create action”, “Create flow module”, or “New folder”. Well, since I’m not a complete beginner I know that I had better create a flow module, since “Create action” will totally bug the LLM out.
Step 4, enter a name. Uh? What??? There’s no step where I set a name, it just created a new flow with the default name “New flow module”. If I want, I can right click that new flow and rename it, but that’s definitely not what step 4 said.
Step 5. Click the return or enter key. Uh what? You don’t click keys, you press them. Yes I’m just quibbling but this is ridiculous. Can we get docs that are actually helpful?
I really want Flows to work out for me. It seems to be the exact thing to be able to help my frontend devs understand the steps they need to follow to use the API effectively. But Flows came out less than a year ago, and the documentation is so outdated. I compared the web version with the Mac version, and both are the same. Every time there is a feature change, you need to fix your documentation. Can’t the LLM go through it and help you fix it?
The next two sections of the article do work (although “input port” and “output port” are very small so might be tricky for a beginner to understand, so suggest you provide a little image or animated gif), but it’s not good that the first section of a beginner article has steps that aren’t exactly what needs to be done.