Skip to main content

Monarch AI

Add the ability to use ChatGPT, GPT4, or other LLMs directly within Monarch

18 comments

Log in to comment and vote

Comments18

  • Nachshon Lenet

    •

    Sep 30, 2024

    If we could bring our own API that would be EPIC

    • rmdashrfv

      Team•

      Oct 1, 2024

      Then it shall be epic!

      Bringing your own API key is the best way to do it because then people would get the full benefit of AI directly in Monarch while still being able to use their services in any and all other apps!

  • L

    •

    Oct 29, 2024

    Is it possible to have built-in prompts?

    Such as YouTube video summarization for the current tab, ask questions on selected files, etc.

    • rmdashrfv

      Team•

      Oct 31, 2024

      My friend, when I debut this next big feature I’m working on make sure your socks are in the ON position, because it just might set them to OFF, if ya know what I mean

  • Mario Dian

    •

    Dec 8, 2024

    Will you be able to use local LLMs? For example, LM Studio lets you create a “ChatGPT” like endpoint which uses the same format and everything.

    • rmdashrfv

      Team•

      Dec 12, 2024

      I’m not familiar with LM studio, but there will be support for local LLMs like Ollama

  • Daniel Mar

    •

    Feb 7, 2025

    A perfect launcher interface to LLMs still doesn’t really exist on mac. The good ones dont let you use whatever (openai compatible) APIs you want.

    For UX inspiration, feel free to look at the chatgpt app “Chat Bar”

    and [Huggingface Chat](https://github.com/huggingface/chat-macOS)

    • Daniel Mar

      •

      Feb 7, 2025

      And please render math and code 🙏
      I just saw this library

      https://github.com/carloskiki/pulldown-latex

      • rmdashrfv

        Team•

        Feb 7, 2025

        I’m going to do my best to have latex out on release! I see it coming up as an issue everywhere!

    • Daniel Mar

      •

      Feb 7, 2025

      And I mentioned the “perfect launcher ai client doesnt exist yet on mac” cause that alone is enough to draw users.

    • Daniel Mar

      •

      Feb 7, 2025

      Other inspiration for AI rust stuff might be the aichat repo on github. it’s a cli AI client

      • rmdashrfv

        Team•

        Feb 7, 2025

        Thanks for these! Can you link me so that I’m 100% sure we’re talking the same repo?

    • rmdashrfv

      Team•

      Feb 7, 2025

      Thank you for this inspiration! 🙌

      • Daniel Mar

        •

        Feb 8, 2025

        Another one in this area that’s under active development is Onit, though it feels a little less polished.

        https://github.com/synth-inc/onit

  • Daniel Mar

    •

    Feb 13, 2025

    Another relatively uncommon (but desirable feature) is Paste Image from Clipboard for multimodal LLMs.

    ChatGPT lets you do this (for 4o):

    Zed editor lets you do this (for claude):

    The Zed code is on github here though idk if it’s particularly useful.

    AIChat lets you do this in their playground (`aichat —serve`). On the cli they only let you manually attach files

    Their js code for the playground looks to be here.

    They let users tell them in config whether the model supports vision

    • Daniel Mar

      •

      Feb 13, 2025

      This is convenient cause on mac we just do ctrl+cmd+shift+4 for clipping screenshots to clipboard

  • olfway

    •

    Aug 25, 2025

    I would love to have a lightweight multiline editor in monarch, so I would be able to write something, modify with with some predefined prompts and some adhoc requests and then copy it to paste in another app or site.

    A lot of apps supports taking already written text and then modify it, but I always forgot how to write multiline text in different apps - sometimes I should press “return”, sometimes “cmd+return”. it would be so easier just to write it in monarch itself.

    May be it could also support ai text completion inline