Jump to content

ChatGPT / DALL-E - OpenAI integrations


Recommended Posts

1 hour ago, sorgel said:

So, basically I'm not sure why we need to duplicate the whole workflow for every new command. Looks like we can stick with editing the original workflow.


You don’t need to duplicate the whole workflow, nor should you edit the original. You should instead make another small workflow with just what you need calling this one. This is explained in the FAQ, under How can I reuse pre-made prompts?

Link to comment
4 minutes ago, vitor said:


You don’t need to duplicate the whole workflow, nor should you edit the original. You should instead make another small workflow with just what you need calling this one. This is explained in the FAQ, under How can I reuse pre-made prompts?

Ah, I see now, sorry, I've misunderstood the original instruction.

Link to comment
On 6/20/2024 at 12:11 PM, lone said:

I made a test example where I try to scroll, and it constantly rolls down

 

 

This behavior is hard-coded in the extension. I'm not sure why @Vero supposed it could be changed with Automation Tasks update.


But we can change it by editing the file chatgpt in the workflow folder. There are several lines behavior: { scroll: 'end' } for different cases.  I don't see the documentation for this thing. But at least changing it to behavior: { scroll: 'start' } will stop infinite scrolling.


What is interesting now is how to hide the first message from the Text view. Personally, I prefer not to see the full prompt, in my case it can be really long.

Link to comment
Posted (edited)
5 minutes ago, vitor said:

You can click the (?) in any object to go to its documentation.


Thanks! Yes, I tried '?" but then i had to find a link to the json description on that page, and that's where I've failed. So `auto` could fit this task more than `start`. 

Edited by sorgel
grammar fix
Link to comment
Posted (edited)
51 minutes ago, sorgel said:

What is interesting now is how to hide the first message from the Text view. Personally, I prefer not to see the full prompt, in my case it can be really long

So the simplest and a bit dirty way to do it is to use  the `replace` in this block

  if (streamMarker)
    return JSON.stringify({
      rerun: 0.1,
      variables: { streaming_now: true },
      response: '…',
      behaviour: { response: 'replace' },
    });


While it ruins the approach where you try to use this workflow for continuous discussion, it makes workflow more usable for one-time prompts

Edited by sorgel
Link to comment
23 hours ago, sorgel said:

So the simplest and a bit dirty way to do it is to use  the `replace` in this block

  if (streamMarker)
    return JSON.stringify({
      rerun: 0.1,
      variables: { streaming_now: true },
      response: '…',
      behaviour: { response: 'replace' },
    });


While it ruins the approach where you try to use this workflow for continuous discussion, it makes workflow more usable for one-time prompts

I tried all possible options, with any replace behavior it scrolls.

Link to comment

Hi, I just installed the ChatGPT / DALL-E workflow, but I am getting an error when trying to prompt ChatGPT.  Error below. Prompting DALL-E works well.

 

[21:32:15.199] ChatGPT / DALL-E[Keyword] Processing complete

[21:32:15.207] ChatGPT / DALL-E[Keyword] Passing output '' to Arg and Vars

[21:32:15.209] ChatGPT / DALL-E[Arg and Vars] Processing complete

[21:32:15.210] ChatGPT / DALL-E[Arg and Vars] Passing output '' to Run Script

[21:32:15.283] ChatGPT / DALL-E[Run Script] Processing complete

[21:32:15.290] ChatGPT / DALL-E[Run Script] Passing output '' to Automation Task

[21:32:15.291] ERROR: ChatGPT / DALL-E[Automation Task] Task not found 'com.alfredapp.automation.core/files-and-folders/path.exists'

 

Workflow version: 2024.11

Alfred version: 5.5 [2257]

Mac OS version: Sonoma 14.5

 

I will appreciate help to torubleshoot the problem.

Link to comment

I feel like I am missing something basic. I would to have text selected in, say, an email I am writing and then hit a keyboard shortcut and be able to use that in a chatgpt session. For example, select text in email program, hit hotkey, type "make this sound more formal" hit return and have alfred replace the text with the response...

When I select the AskChatGPT universal action, it sends the text directly to ChatGPT, which I don't find helpful.

Link to comment

Welcome @Steev28,

 

I can see the sense in altering the message before sending specifically when using a Universal Action where the text you want to use may not have the proper context. I’m experimenting with it and this version does it. It won’t auto-paste the result, as that’s a bit outside the scope of the workflow but on the result you can ⌥↩ to copy it to the clipboard.

 

This is a first draft and may change or be reverted before release.

Link to comment

Seeing an issue in the debugger when I try and use this workflow:
 

Code 1: /Users/robdeskoski/Library/Application Support/Alfred/Alfred.alfredpreferences/workflows/user.workflow.8DF3C6CF-EE0A-4AFC-8C1D-2E383942484A/chatgpt: execution error: Error: SyntaxError: JSON Parse error: Expected ']' (-2700)

 

API Key seems fine, typing in as per the examples.

Link to comment

@ged12345 When reporting issues, please include your exact installed versions of:

  • The Workflow.
  • Alfred.
  • macOS.

In addition to:

  • The debugger output. Perform the failing action, click “Copy” on the top right and paste it here.
  • Details on what you did, what happened, and what you expected to happen. A short video of the steps with the debugger open may help to find the problem faster.

These requirements are listed in the FAQ (second post on the thread).

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...