[ChatGPT Quick Actions] Custom API Base URL setting (and model switcher based on /v1/models)

[FallenChromium]

Extension

https://www.raycast.com/alanzchen/chatgpt-quick-actions

Description

It would be nice to be able to use Ollama or other OpenAI-compatible endpoints (LiteLLM proxy being a popular example) with this extension. This setting is present in the "ChatGPT" extension (brought up to make an example of the UX), I use this extension to chat with local LLMs with comfort, but since "ChatGPT Quick Actions" provides wonderful "Execute" and "Transform" commands that are not present in the mentioned extension, I would love to set it up the same way.

Who will benefit from this feature?

Addition of the Base URL setting, the change to make API key optional and changing the model switcher to make it universal will provide tinkerers and privacy-focused people with means to use alternative LLM providers with this extension, which is cool if you have usecases that do not require "big" LLMs (one example being code refactors - MacBooks with 32GB RAM can run models that are capable of this without slowing down the system, this greatly improves privacy, latency and sometimes even accuracy)

Anything else?

No response

[raycastbot]

Thank you for opening this issue!

🔔 @alanzchen @one-data-cookie @JonathanAquino you might want to have a look.

💡 Author and Contributors commands

The author and contributors of alanzchen/chatgpt-quick-actions can trigger bot actions by commenting:

• @raycastbot close this issue Closes the issue.
• @raycastbot close as not planned Closes the issue as not planned.
• @raycastbot rename this issue to "Awesome new title" Renames the issue.
• @raycastbot reopen this issue Reopens the issue.
• @raycastbot assign me Assigns yourself to the issue.
• @raycastbot good first issue Adds the "Good first issue" label to the issue.
• @raycastbot keep this issue open Make sure the issue won't go stale and will be kept open by the bot.

[one-data-cookie]

Using open-source models like Ollama myself, I love the idea! Unfortunately, I don't have the capacity to implement this at the moment.

@alanzchen @JonathanAquino?

[alanzchen]

I don't have any bandwidth for this either.

[JonathanAquino]

I don’t think so - I definitely saw an OpenAI library hardcoded into the commands. So I don’t see how this could work with Anthropic, etc., without some big changes.

…

________________________________ From: Alan Chen ***@***.***> Sent: Thursday, September 19, 2024 2:50:09 PM To: raycast/extensions ***@***.***> Cc: Jonathan Aquino ***@***.***>; Mention ***@***.***> Subject: Re: [raycast/extensions] [ChatGPT Quick Actions] Custom API Base URL setting (and model switcher based on /v1/models) (Issue #14540) I don't have any bandwidth for this either, but won't the base URL option allow you to use whatever model you want? — Reply to this email directly, view it on GitHub<#14540 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAAWBMDJ3GOVWQX2UNUIZJTZXNBJDAVCNFSM6AAAAABOORU7AWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNRSGI3DGNBYGQ>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[raycastbot]

This issue has been automatically marked as stale because it did not have any recent activity.

It will be closed if no further activity occurs in the next 10 days to keep our backlog clean 😊