At the moment, this is a quick and dirty prototype of a ChatBot UI to use GPT-3 in a similar way to ChatGPT by API calls. Keeping the API calls on the server side.
The project's goal aren't fully defined and may drastically change in the future. At the moment it's meant for personal usage. But I may end up making my server side code not depend on this specific front so that it could be used for instance as a Discord bot.
It currently serves as an AI assistant and support markdown. Ideally the client should be able to tweak the prompt to change the AI's emotions or purpose (i.e. fictional character impersonation)
# Install requirements with
pip install -r requirements.txt
Create a config.py file and write
API_KEY = "<OPEN AI API KEY HERE>"
Then run main.py and open localhost:5000
- Markdown escape
- Connect with DALL-E 2
- Connect with WolframAlpha
- Save conversations
- Retrieve conversations
- Delete conversations
- Count tokens instead of characters
- Display remaining OpenAI credits (If Possible)
- Better short term memory1
- Long term memory
- Create "Character" system
- Save character
- Retrieve character
- Delete character
- Characters should have a name
- Characters should have a role
- Characters may have a personality
- Characters may have an emotion (maybe dynamic)
- Characters may have long term memory1