Skip to content

A Remake of the Google Gemini Fake Demo, Except Using GPT-4 and It's Real

Notifications You must be signed in to change notification settings

mdwoicke/sagittarius

 
 

Repository files navigation

Sagittarius

What is this? A GPT-4/Gemini Voice/Video Exploration Tool!

Do you have an API key from either OpenAI or Gemini? You can use the tool online! No need to install anything.

See below for more context:

how to build

  • clone this repo, cd into it
  • duplicate .env.example and name the copy .env
  • fill out the VITE_OPENAI_KEY= value with your OpenAI api key. you must have access to the gpt-4-vision-preview model
    • you can also try out the Gemini API if you have a key -- fill out VITE_GEMINI_KEY in the same .env
  • then, run:
  • npm install
  • npm run dev
  • the demo will be running at http://localhost:5173

note: the in-browser speech recognition works best in Google Chrome

TODO

  • allow input of API keys as <input> on the page
  • deploy frontend to site i.e. sagittarius.greg.technology via vite+github actions
  • enable streaming output..!
  • make new video with 3) streaming output / comparison
  • enable selection of dictation language
  • make new video with 1) uses of repo in the wild / forks 2) UI improvements
  • add allcontributors bot
  • add dependabot

About

A Remake of the Google Gemini Fake Demo, Except Using GPT-4 and It's Real

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • TypeScript 63.3%
  • JavaScript 18.6%
  • CSS 9.1%
  • HTML 9.0%