- Make it easy to create and deploy applications that use AI
- It is more fun and rewarding to have something actually finished, deployed, and useable as a starting point that you can then build on top of.
- You will better discover what you actually need by building MVPs in this way.
- This guide and code is not meant for large scale production. It is meant to be a starting point for people who want to learn how to use AI in their applications. Creating an application for yourself and or a few people.
- It is optimized to be quick and easy to start, deploy, and learn from.
-
Get an open AI key. set a low budget so you don't get surprised by the bill. You can always increase it.
-
fork this repo
-
clone your fork
-
create a virtual environment
-
install the requirements using
pip install -r requirements.txt -
run the app using
streamlit run app.py
- see https://github.com/features/codespaces for detailed instructions
- Fork this repo
- go to settings -> codespaces -> click set up prebuild
- in the repo, under code, click "start codespace"
- to launch from command line run
python -m streamlit run app.py
- see RPG app for an example of how to customize the application
- Estimates cost of API requests so you don't get surprised by the bill
- Includes conversation history and allows you to change how much if any to use
Put your application on the web so you can easily access it
- Use chatgpt and or github copilot to help you write and troublelshoot your code
- when your code gets to be long, consider some refactorings:
- extract methods
- separate display logic from business logic
- Use gpt3.5-turbo when you can instead of gpt-4. It is much cheaper.