quantumfrost00 Posted September 2 Share Posted September 2 This is by far one of the best and probably my favorite workflow for alfred. I am a medical student and while doing anki, I need my questions answered in a specific context. I used to use google but it wasn't sufficient as it always gives sources that generalise the answer. I switched to Chatgpt but it was always a hassle to open a browser window, login again (sometimes), and wait for interface to load up (not to mention the occasional errors while generating the repsonse) . I came across this by chance. After configuring this workflow and finally having it up and running, I must say this is a literal quality of life improvement. Please continue maintaining it. Suggestion: An option to make the text in chat interface larger or smaller will be a very nice and welcome addition since it becomes difficult to read small text in certain alfred themes. Link to comment
zeitlings Posted September 2 Author Share Posted September 2 Hey @quantumfrost00 Thanks for the kind words! 4 hours ago, quantumfrost00 said: Suggestion: An option to make the text in chat interface larger or smaller will be a very nice and welcome addition since it becomes difficult to read small text in certain alfred themes. Note that you have some control over the text size: On 8/17/2024 at 2:08 PM, zeitlings said: At the bottom of the configuration you can find a section called "Accessibility" where the text size can be changed. The available sizes are enforced by Alfred's text view and cannot be customized further. Generally, the look of the text view is determined by your current theme, i.e. changes in color or font size there will be reflected in the Ayai chat window. Unfortunately, ad hoc changes to the theme of a specific object (such as the text view used for the chat window), without affecting the theme everywhere else, is currently not possible. quantumfrost00 1 Link to comment
jpomykala Posted September 11 Share Posted September 11 (edited) Hey @zeitlings, I've configured the plugin, set up the ChatGPT API Key, and I've configured a hotkey for "Inference Actions," but when I select any option, Alfred just closes and nothing happens. I tried with the selection option and clipboard option; I also tried to change the models, but still nothing. 🤔 I'm pretty sure the API Key is correct as I used it in other projects. PS Cool theme, do you mind sharing it also? 😄 Edited September 11 by jpomykala Link to comment
zeitlings Posted September 11 Author Share Posted September 11 (edited) Hey @jpomykala Could you share the output of Alfred's Debugger Utility? Also, keep the behaviour and prerequisites in mind: On 8/21/2024 at 10:46 AM, zeitlings said: The results will be either streamed, pasted or copied to the clipboard. The default behaviour can be set via the Generate configuration option or selected en passant via the key modifier combinations. Ayai will make sure that the frontmost application accepts text input before streaming or pasting anything, and will simply copy the result to the clipboard if it does not. This requires accessibility access, which you may need to grant in order to use inference actions. If you're using the Hotkey or Universal Action in a text editor, the results may be streamed or pasted, depending on your configuration. Otherwise the result will (at the moment silently) be copied to the clipboard. In that case, try to just paste with cmd+v after a while. 1 hour ago, jpomykala said: I'm pretty sure the API Key is correct as I used it in other projects. If normal conversations via the chat window work, you can be sure that the API Key is valid. 1 hour ago, jpomykala said: PS Cool theme, do you mind sharing it also? 😄 Sure, I'll export and upload it somewhere in a bit. Available: here. Edited September 11 by zeitlings Themes Link to comment
quantumfrost00 Posted Tuesday at 03:30 AM Share Posted Tuesday at 03:30 AM Hi, can you add the ability to make my own custom inference actions. Or ability to execute an inference based on my prompt ? Link to comment
quantumfrost00 Posted Tuesday at 12:23 PM Share Posted Tuesday at 12:23 PM 8 hours ago, quantumfrost00 said: Hi, can you add the ability to make my own custom inference actions. Or ability to execute an inference based on my prompt ? Also I don't know if it's a bug or not. Infrerence actions don't work in apple mail or thunderbird where I really need it. It works normally in the Apple Notes app. Is there any setting in alfred/workflow that I need to change or is it normal ? Link to comment
zeitlings Posted Tuesday at 03:26 PM Author Share Posted Tuesday at 03:26 PM 11 hours ago, quantumfrost00 said: Hi, can you add the ability to make my own custom inference actions. Or ability to execute an inference based on my prompt ? Hey, you already can add custom inference actions by modifying the `actions.json` in the workflow folder. See also the discussion here #6. 2 hours ago, quantumfrost00 said: Also I don't know if it's a bug or not. Infrerence actions don't work in apple mail or thunderbird where I really need it. It works normally in the Apple Notes app. Is there any setting in alfred/workflow that I need to change or is it normal ? That's "normal," and I've already noticed it in Apple Mail. The problem is that Apple is hiding the fact that you're currently in a view that accepts text. This prevents the safety feature (stream or paste iff the frontmost app's selected view accepts text) from detecting that you're actually typing or editing text in Apple Mail, leading to an opt-out. I haven't found a workaround yet. The only reliable method at this time is to ignore the safety check altogether for specific apps. Naturally, I'd like to avoid that, but I'm considering making it possible through an environment variable that needs to be manually populated. I find Apple Mail's issues here mildly frustrating myself 😅. quantumfrost00 1 Link to comment
quantumfrost00 Posted Tuesday at 03:43 PM Share Posted Tuesday at 03:43 PM 12 minutes ago, zeitlings said: Hey, you already can add custom inference actions by modifying the `actions.json` in the workflow folder. See also the discussion here #6. That's awesome. I didn't know about it. 13 minutes ago, zeitlings said: That's "normal," and I've already noticed it in Apple Mail. The problem is that Apple is hiding the fact that you're currently in a view that accepts text. This prevents the safety feature (stream or paste iff the frontmost app's selected view accepts text) from detecting that you're actually typing or editing text in Apple Mail, leading to an opt-out. I haven't found a workaround yet. The only reliable method at this time is to ignore the safety check altogether for specific apps. Naturally, I'd like to avoid that, but I'm considering making it possible through an environment variable that needs to be manually populated. I find Apple Mail's issues here mildly frustrating myself 😅. Thankyou, it would be very helpful. Thunderbird and a couple other apps also exhibit the same behavior. Hopefully the next update fixes this. Rooting for the alpha version 7 🙌 Link to comment
Recommended Posts
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 accountSign in
Already have an account? Sign in here.
Sign In Now