in reply to Double Tap

mastodon - Link to source

Pitermach

@fireborn @payown I use Vollama here. It takes some time to start, but once it gets going it's a very simple UI. It also supports most major models including local ones. github.com/chigkim/VOLlama

GitHub - chigkim/VOLlama: An accessible chat client for Ollama

An accessible chat client for Ollama. Contribute to chigkim/VOLlama development by creating an account on GitHub.
GitHub
@aaron @Michael Babcock
in reply to Double Tap

mastodon - Link to source

Pitermach

@kevinrj @fireborn @payown Yes. I don't know if you ever used any apps that require your own OpenAI API key. If not, you have to go to platform.openai.com, log in with your account, create a new key and top up your account. Then in Vollama press Command+Shift+A or select API settings from the chat menu and you can enter the key you created earlier. For topping up, this is a separate thing from a ChatGPT Plus subscription, there is a 5 dollar minimum but it's very likely using it this way will be cheeper because your credits are valid for a year and one request usually costs only a few cents. And there are many other apps you can plug that key into, for example VOCR from the same developer which will let you get AI image descriptions of anything on the screen with 1 keystroke.
@Kevin R Jones @aaron @Michael Babcock
in reply to Double Tap

mastodon - Link to source

Pitermach

@kevinrj @fireborn @payown You can't download any of the GPT models because they're simply way too big to run for a typical computer. But if you want to play with other local models, no it's not just meta ones that you can get for ollama. Best place to look is on their website but you can find things like Gemma which is a small model from Google, Mistral AI or even small versions of Deepseek which are all good at different tasks. I'm also only getting into the local models that are available myself but this should be a good start to explore this :)
@Kevin R Jones @aaron @Michael Babcock
⇧