Don't know how to implement a function? Let Universal API help you on the fly!

This is a non-serious library that can implement any functions on the fly using LLMs. To use it, simply call a function with a descriptive name and parameters, and it will be defined, implemented and called at runtime. Generated functions are cached so you don't pay for things that are already implemented.

(Note: this repo is based on my repo at https://github.com/pycui/universal-api and adopted to Cohere API.)

api = UniversalAPI()
# Start to call arbitrary functions 
print(api.sort([3, 2, 1])) # returns [1, 2, 3]
print(api.sort([4, 3, 2, 1])) # returns [1, 2, 3, 4] using cached implementation
print(api.sort([1, 2, 3], reverse=True)) # returns [3, 2, 1]
print(api.add(1, 2)) # returns 3
print(api.reverse('hello')) # returns 'olleh'
api.fizzbuzz(15) # prints the fizzbuzz sequence up to 15
api.print_picachu() # prints an ASCII art of Picachu

This library will execute unverified code in your local machine. It's NOT safe to run in production (or really, any serious environment).

By default this library uses OpenAI GPT API. You can modify environment variables like OPENAI_BASE_URL to use other LLM endpoints (e.g. Anyscale Endpoint), which allows you to run on other models like the LLaMa series. It's also possible to use local LLMs.