I’ve been building and maintaining LLM-God, a desktop LLM prompting app for Windows, built with Electron. It allows you to ask one question to multiple LLM web interfaces at once and see all the returned answers in one place. If you hate tabbing through multiple browser tabs to ask multiple LLM's the same question, this project is the antidote for that.

It is using JavaScript to inject the global user prompt into the HTML DOM bodies of the individual browser views, which contain the webpages of the different LLM's. When the user clicks Ctrl + Enter, a message is sent to the main app which tells the individual pages to programmatically click the "send" button. The communication using IPC is also happening when the user tries to add more LLM browser views to the main view.

The challenging part for me was to come up with the code for allowing the individual LLM websites to detect user input and the clicking of the send button. As it turns out, each major LLM providers often change the makeup of the HTML bodies for some reason, causing the code to break. But so far, the fixes have been manageable.

Key features:

Starts with a default of Perplexity, ChatGPT, and Gemini, with the option to add more LLM's like Grok, Claude, and DeepSeek.

Responsive, keyboard-friendly interface

Ability to add, edit, and delete your own custom prompts that you can inject into the global prompt area. If you have custom prompting templates that you like to use, this can help with that!

Feedback is welcome here, on GitHub: https://github.com/czhou578/llm-god

YouTube Demo: https://youtu.be/YxqWUp0Wmi0