MacOS AI Assistant: Your Questions Answered Instantly! (NodeJS + OpenAI Vision/Whisper/TTS tutorial)

[@willwang22]

When I install the npm method and cloned the git. The application started successfully and is running, but. it says that "there was an error transcribing your recording". What could possibly be the problem? I've enabled the microphone access to the terminal but it doesn't seem to work.

[@willwang22]

Update: I've changed the input method to text then found out that it says "there was an error calling the OpenAI Vision API". Where should I possibly diagnose to reach for the problem?

[@mybalilife365]

Hey, great example. I just get a small error. OpenAI has changed the name (NotFoundError: 404 The model `gpt-4-vision-preview` has been deprecated) Where is that model name?

[@darioalejandro180]

When attempting to use the app, I repeatedly encounter an error message stating, 'There was an error transcribing your recording.' I would greatly appreciate any troubleshooting steps or guidance you could provide 🙏

[@michaellevey3681]

Hey, great example. I just get a small error. OpenAI has changed the name (NotFoundError: 404 The model `gpt-4-vision-preview` has been deprecated) So to fix you have to update the model to "gpt-4o-mini" and you have to pay $10 into your openai account, otherwise you get openai "too many requests" for free account

[@shylov1740]

we don’t have to pay for have an API key of OpenAi?

[@jackculpan]

HN sent me here love it

[@ralfelfving]

Glad you found it! :)

[@barskratl5833]

I tested on Windows PC but I couldn't handle in any way. Looking forward to see windows version of it! Thank you for this amazing idea!

[@ralfelfving]

Alright, thanks for letting me know! Hoping someone can pick up the Windows version of it, without having run Node/Electron on Windows I still think not much need to change in the script.

[@rolandcucicea6006]

A lot of people ask me why I own a Mac, and it's comments like these that make me chuckle a bit. I use a windows computer for my office, but once you're locked in a mac/unix environment for long enough it's hard to let go. The only thing I regret is not buying a 32GB RAM machine instead of my 16GB so I could run more powerful local LLMs. Can I game locally? No. But I can do everything else faster.

[@kris-fo5wm]

Very nice. Thanks for explaining.

[@ralfelfving]

Glad it was helpful!

[@kate-pt2ny]

今天碰巧看到这个视频，还发现了博主的最新关于这个产品的app，恭喜，干得不错

[@PROPHAMMER]

super rad dude!

[@mariomancev78]

Does this work with a 3.5 api key? Or do you need a gpt 4 api key?

[@ralfelfving]

I don't think API keys are scoped to models, so there's no such thing as 3.5 or 4 API keys. Any OpenAI API key should work with any model, unless they have restrictions of new models while in beta -- but that'd likely be scoped by account.

[@mariomancev78]

@@ralfelfving I see, thanks for your reply. I'll look into that!

[@wobsoriano]

Love this! I'll probably clone this and use it on my next local conference demo :)

[@ralfelfving]

Glad to hear it was interesting!

[@TomsDone]

Looks great. Does it work with text prompts instead of audio converted to text? Sorry, I didn't watch the whole video because I think the answer is no. Thanks either way.

[@ralfelfving]

No, it's currently audio only because i mainly built it for my own use (and i sit alone, and speak faster than I type).

[@petrbrabec7856]

If you check the video, you'll understand it's pretty simple to add that functionality :)

[@massibob2004]

Good job ❤ Can we have the same on Windows ?😊

[@ralfelfving]

Thanks. Not by me unfortunately, don't have windows machine. I'll extend the functionality, but focus on mac

[@ralfelfving]

@TomsDone the text input was added today.

[@musicsavestheworld1383]

Ali G making some cool AI tools