skaplanofficial / raycast-promptlab Goto Github PK

I tried to run Commands with this extension

My configuration is as follows:

{ "model": "gpt-3.5-turbo", "messages": [{"role": "user", "content": "{prompt}"}]}

The problem is, that the request is sent, I am billed from OpenAI, but the extension says it got no responses.

This could improve performance and accuracy when dealing with large files or web pages.

How to use {{input}}, {{clipboardText}}, and {{selectedText}} at the same time???

Hello! I am using the OpenAI endpoint.
When I run TechCrunch News or Summarize GitHub trending, I get responses like

As an AI language model, I do not have the ability to analyze today’s trending GitHub repositories as they change frequently, but I can provide a summary of the current trending repositories based on the given link.

It seems the the content from the URL is not being parsed into the prompt.

Here are the steps to reproduce:

1. Open "My PromptLab Commands"
2. With any command, open "Edit command" menu
3. Scroll down and change the modal from the current selected to any other
4. Save the command

Expected result: the command is correctly saved with the new model used, each command run shows different results because it uses different model.
Actual result: the command is not saved, if you go back to the "edit command" menu the model is just as it was before, the results of the command run are similar to the previous.

Steps to reproduce:

• Install latest version from store (1.2.1).
• Configure with OpenAI example here and a valid OpenAPI key.
• Select PromptLab > New Chat
• Enter any text.
• Run the command.

Results

"Bad Request"

E.g. by separating prompt and input

When I run the summarise github trending command, it pulls out the first entry but outputs this to the terminal. I am using OpenAI endpoint.

11:28:03.889 Failed to get JSON from model output
11:28:03.889 Warning: Cannot update a component (`CommandResponse`) while rendering a different component (`CommandResponse`). To locate the bad setState() call inside `CommandResponse`, follow the stack trace as described in https://reactjs.org/link/setstate-in-render
    at CommandResponse (/Users/hsai002/.config/raycast/extensions/promptlab/search-commands.js:16447:11)
    at Nr (/Applications/Raycast.app/Contents/Resources/RaycastCommands_RaycastCommands.bundle/Contents/Resources/api/node_modules/@raycast/api/index.js:6:2490)
    at ray-navigation-stack
    at _o (/Applications/Raycast.app/Contents/Resources/RaycastCommands_RaycastCommands.bundle/Contents/Resources/api/node_modules/@raycast/api/index.js:6:2088)
    at Suspense
    at wr (/Applications/Raycast.app/Contents/Resources/RaycastCommands_RaycastCommands.bundle/Contents/Resources/api/node_modules/@raycast/api/index.js:5:2543)
    at ray-root
    at ti (/Applications/Raycast.app/Contents/Resources/RaycastCommands_RaycastCommands.bundle/Contents/Resources/api/node_modules/@raycast/api/index.js:6:3000)```

Any plans to support Azure OpenAI endpoints? I've tried it with what seems to be the correct settings, but no luck yet.

I'm having difficulty debugging what's actually happening. All I've been able to see is the chat gets named Unauthorized.

Azure OpenAI shouldn't be much different from regular OpenAI from what I can tell.

Currently the Raycast API makes the following models available via the Model option:

type Model =
  | "anthropic-claude-haiku"
  | "anthropic-claude-opus"
  | "anthropic-claude-sonnet"
  | "openai-gpt-3.5-turbo-instruct"
  | "openai-gpt-3.5-turbo"
  | "openai-gpt-4"
  | "openai-gpt-4-turbo"
  | "perplexity-sonar-medium-online"
  | "perplexity-sonar-small-online"
  | "llama2-70b"
  | "codellama-70b-instruct"
  | "mixtral-8x7b";

This functionality was introduced in v1.71.0.

Error:

Error: Model is not supported
    at Rr (/Applications/Raycast.app/Contents/Resources/RaycastCommands_RaycastCommands.bundle/Contents/Resources/api/node_modules/@raycast/server/index.js:25:230)
    at Ye (/Applications/Raycast.app/Contents/Resources/RaycastCommands_RaycastCommands.bundle/Contents/Resources/api/node_modules/@raycast/server/index.js:24:6249)
    at ht (/Applications/Raycast.app/Contents/Resources/RaycastCommands_RaycastCommands.bundle/Contents/Resources/api/node_modules/@raycast/server/index.js:24:6374)
    at Immediate.<anonymous> (/Applications/Raycast.app/Contents/Resources/RaycastCommands_RaycastCommands.bundle/Contents/Resources/api/node_modules/@raycast/server/index.js:24:6139)
    at process.processImmediate (node:internal/timers:478:21)

When I launch a command, the error appear at the bottom left.

The error do not prevent the command from working, but it is a little annoying.
I am using a local model hosted with Ollama. I have tried to configure the model using both extension settings and the model manager.
At this moment, I am not sure what the extension is trying to do, but my command works fine.

I recently came across this library called Jan.ai, which allows you to run offline models such as Mistral Instruct, which is really powerful. Apparently, this tool has the same API as OpenAI.

Is it possible to use it with this extension for Raycast? If so, is there any resources I could read about this? And if you need my help, please do tell.

This is an example of how I can access the model using their API which is the same as OpenAI's:

curl http://localhost:1337/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer EMPTY" \
  -d '{
     "model": "mistral-ins-7b-q4",
     "messages": [{"role": "user", "content": "Say this is a test!"}],
     "temperature": 0.7
   }'

Reply:

{"choices":[{"finish_reason":null,"index":0,"message":{"content":" I understand that you have asked me to say that \"this is a test.\" Here is that statement for you: \"This is a test.\" Is there anything specific you would like me to do with this test, or is it simply for my understanding that we are conducting a test? Let me know if there's anything else I can assist you with.","role":"assistant"}}],"created":1705970517,"id":"tWF3q25T3yaeF96K2bVF","model":"_","object":"chat.completion","system_fingerprint":"_","usage":{"completion_tokens":72,"prompt_tokens":14,"total_tokens":86}}% 

Hey! So, I'm having some issues when trying to use the Claude3 models with PromptLab, it's not playing nice with the latest version of Anthropics' 'messages' API. The problem is that the new API needs an extra header, "anthropic-version", in the request. Since the extension settings doesn't include a space to specify additional headers, I can't make successful API calls.

Here's an example of what the request should look like:

curl https://api.anthropic.com/v1/messages
--header "x-api-key: $ANTHROPIC_API_KEY"
--header "anthropic-version: 2023-06-01"
--header "content-type: application/json"
--data
'{
"model": "claude-3-opus-20240229",
"max_tokens": 1024,
"messages": [
{"role": "user", "content": "Hello, world"}
]
}'

Could you please help update the extension to include this new header? Thanks a bunch in advance!

Hi,

I installed your extension earlier today and it's fantastic. It was working fine but now when I go to 'My PromptLab Commands' I get the following error:

TypeError: command.acceptedFileExtensions?.split is
not a function
SearchCommand:search-commands.tsx:27:2

24: import path from "path";
25: 
26: export default function SearchCommand(props: { arguments: { commandName: string; queryInput: string } }) {
27:   const { commandName, queryInput } = props.arguments;
     ^
28:   const [commands, setCommands] = useState<Command[]>();
29:   const [targetCategory, setTargetCategory] = useState<string>("All");
30:   const [searchText, setSearchText] = useState<string | undefined>(

SearchCommand:search-commands.tsx:27:2
Nr:index.js:6:2490
    at ray-navigation-stack
_o:index.js:6:2088

I installed a few batches of commands from the store. Most worked well. One issue I had, almost certainly unrelated, was when summarising PDF with filename Walter Murch - In the Blink of an Eye Revised 2nd Edition (2001, Silman-James Pr).pdf, it didn't like it and gave me the error:

Error: ENOENT: no such file or directory, Istat '/Users/
Alex/Downloads/Walter Murch - In the Blink of an Eye
Revised 2nd Edition (2001'
Obiect.IstatSync:node:fs:1569:3

![Finder - Downloads - 13-05-2023 - 20 14 22@2x](https://github.com/SKaplanOfficial/Raycast-PromptLab/assets/11060179/c47d2ea3-d6ad-411e-8d39-747765fa2813)

Renaming it to test.pdf worked and then everything seemed fine after.

The last command I tried to run was 'summarise spoken audio' which gave me an error that unfortunately I didn't capture.

Thanks,

Alex

Hi, PromptLab is having trouble accessing my calendar events. When I run the prompt 'Today's Agenda' that comes with the extension, I get the following error:

Error: Command failed with exit code 1: osascript -e use framework "EventKit"
      property ca : current application
      
      set eventStore to ca's EKEventStore's alloc()'s init()
      eventStore's reset()
      eventStore's requestAccessToEntityType:((get ca's EKEntityMaskEvent) + (get ca's EKEntityMaskReminder)) completion:(missing value)
      delay 0.1
      
      set startDate to ca's NSDate's |date|()
      
      set calendar to ca's NSCalendar's currentCalendar()
      set dateComponents to ca's NSDateComponents's alloc()'s init()
      dateComponents's setDay:1
      set endDate to calendar's dateByAddingComponents:dateComponents toDate:startDate options:(ca's NSCalendarMatchStrictly)
      
      set remindersPredicate to eventStore's predicateForIncompleteRemindersWithDueDateStarting:startDate ending:endDate calendars:(missing value)
      set upcomingReminders to eventStore's remindersMatchingPredicate:remindersPredicate
      set theRemindersData to {title, dueDate} of upcomingReminders
      
      set theReminders to {}
      repeat with index from 1 to (count of upcomingReminders)
        set eventTitle to (item index of item 1 of theRemindersData) as text
        set eventDueDate to item index of item 2 of theRemindersData
        
        set dueDateFormatter to ca's NSDateFormatter's alloc()'s init()
        (dueDateFormatter's setDateFormat:"MMMM dd, YYYY 'at' HH:mm a")
        set eventDueString to (dueDateFormatter's stringFromDate:eventDueDate)
        
        set reminderInfo to eventTitle & " on " & eventDueString
        copy reminderInfo to end of theReminders
      end repeat
      
      return theReminders 
932:948: execution error: Can’t get title of missing value. (-1728)

      property ca : current application
      
      set eventStore to ca's EKEventStore's alloc()'s init()
      eventStore's reset()

I tried creating my own prompt using {{weekEvents}} but it returns nothing, even though there are plenty of events on my calendar.

Any help would be appreciated! Thank you.

The repo has tremendous value but it would be great if you have some examples in video format. Thanks!

Hey, I just tried this extension and it is so good! Thank you very much! Is there any way that I can configure the AI (I'm using OpenAI API with my own key), to respond in German? Currently, even when I feed the AI with German content, it responds in English.

I've been enjoying Raycast-PromptLab since discovering it this morning, until this afternoon's Raycast update (1.51.0) which officially released Raycast AI... 😅

No matter which prompt I use, I get the same message, even after subscribing to the Raycast Pro membership.

Thank you for your work on Raycast-PromptLab!

Reason: No executable JS file found
Domain: commands
Time: 01:32:04.042
Recovery: Build the extension before usage

Hi, how can I edit the Summarize Current Tab or Summarize Last Email commands so that they use different default apps (Edge instead of Safari, Outlook instead of Mail)?

Discussed in #29

#popclip extension snippet to replace the selected text with output from Raycast AI
name: Replace With AI Response
icon: AI
applescript: do shell script "open -g raycast://extensions/HelloImSteven/promptlab/search-commands?arguments=%7B%22commandName%22:%22Ask%20AI%20%26%20Replace%20Selection%22%7D"

See VNGenerateAttentionBasedSaliencyImageRequest

Is there any way to make the main action be pasting the response to the active app instead of copying to the clipboard? Thank you.

Would be great to better emulate the "AI Chat" UI that can be found in Raycast pro..

Hi, PromptLab has been working fairly well for me when used without placeholders (anything with {{}}) However as soon as I include terms like {{selectedText}} the prompts will show "loading response" indefinitely. I'm using my personal (not Raycast( openAI endpoint - is there any way to diagnose what exactly the issue is?