This is an AI dashboard that I created for myself. It's hosted at ai.esezen.me. You can play with it if you bring your own Open AI API key. The keys are stored at localStorage and not shared with any other service than Open AI itself. At the moment there are 3 things you can do:
The normal ChatGPT like interface. Let's you use GPT4
Prompt to image
Audio file to text. I thought this might be useful for converting voice memos to notes
If you are reading this, you probably know how to run this locally, but in case you don't, it's First, run the development server:
npm run dev
# or
pnpm dev
Open http://localhost:3000 with your browser to see the result.
- Clean up the code
- The reducer logic is especially bad
- Authentication using Supabase
- Update other parts to use data from DB instead of localStorage
- Save things to an actual database
- I currently same images to localStorage which hits the quota after a few images
- Maybe move the Open AI calls to a Lambda. (Vercel doesn't allow functions to run longer than 10 seconds on free tier)
- Connect to a locally running modal (Ollama)
- Import / export from localStorage
- Stable diffusion support for images
- Upload PDF
- Upload image