StreamlinedGPT

StreamlinedGPT is a Python library designed to simplify interactions with leading AI models. It provides a unified interface to seamlessly integrate and experiment with various models, removing the complexity of dealing with multiple APIs. The library also streamlines the process of granting AI access to tools, empowering users to leverage AI capabilities in a more dynamic and intuitive way, without the need to handle verbose syntax or repetitive boilerplate code.

Whether you're an individual innovator or an enterprise team, StreamlinedGPT is the perfect solution for experimenting with AI models, tooling workflows, and custom prompts. It's design prioritizes ease of use, aiding rapid prototyping and fostering creativity. By reducing development friction, it enables users to focus on building meaningful applications rather than wrestling with technical details.

The GitHub page can be accessed here https://github.com/XavierEgan/StreamlinedGPT

Why use StreamlinedGPT?

StreamlinedGPT is designed to simplify the syntax of existing solutions, like OpenAI's Python library, enabling rapid prototyping of tools, prompts, and experiments.

It addresses the lack of an easy, quick way to experiment with model tooling by eliminating boilerplate code and reducing complexity.

Below is a comparison of StreamlinedGPT and OpenAI's library.

import StreamlinedGPT

chatbot = StreamlinedGPT.Assistant(
    StreamlinedGPT.Openai_Text_Adaptor()
)

chatbot.chatloop()

from openai import OpenAI
client = OpenAI()

message_history = []

while True:
    user_input = input("\033[94mUSER > \033[0m")

    if user_input == "exit":
        break

    message_history.append(
        {
            "role": "user",
            "content": user_input
        }
    )

    response = client.responses.create(
        model="gpt-4o-mini",
        input = message_history
    )

    message_history.append(
        {
            "role": "assistant",
            "content": response.output_text
        }
    )

    print(f"\033[92mASSISTANT > \033[0m{response.output_text}")

Limitations

It currently only has OpenAI's models. However adding more models is extremely easy due to the structure of the project.

It is more limited compared to OpenAI's API. It lacks 4o's real time capabilities, structured outputs and streaming. However, this library is not intended to be used in a final product. Its intention is to be a tool for fast prototyping, trading off functionality for quick development.

Quick Start

The below code gives a simple example of how the library can be used

import StreamlinedGPT

chatbot = StreamlinedGPT.Assistant(
    StreamlinedGPT.Openai_Text_Adaptor()
)

chatbot.chatloop()

The Assistant class takes in an Adaptor. In this example the Openai_Text_Adaptor was used. Currently the library only has an openai adaptor, however this will be expanded in the future.

Tool use is quite simple with StreamlinedGPT. A tool is a python function that the ai model can call. Below is an example

import StreamlinedGPT

chatbot = StreamlinedGPT.Assistant(
    StreamlinedGPT.Openai_Text_Adaptor()
)

def print_to_console(s):
    print(s)

chatbot.add_tool(
    tool=StreamlinedGPT.Tool(
        function=print_to_console,
        name="print_to_console",
        description="prints something to the console",
        arguments=[
            StreamlinedGPT.Argument(
                name="s", 
                type="string", 
                description="the string to print"
            )
        ]
    )
)

chatbot.chatloop()

A tool is defined by making an instance of the Tool class, and passing it into the add_tool method of an Assistant.

Documentation

class Argument

A class that represents an argument for a tool.

constructor

class Assistant

A class that provides methods for having chat-like conversations with AI models.

constructor

method send_message

method send_message_without_history

method add_tool

method chatloop

Starts an interactive chat loop in the console.

class Tool

A class that represents a tool that can be used by the AI model.

constructor