This project tries to be a free library to create a python assistant with GPT4 and other providers like Bing Search, Watson Speech to Text, Bing Speech to Text, Google Text to Speech and more.

It's a python library that can be used to import in other projects, without any API key, and with free internet partners.

There is a pytest file that could be used to test the library. Anyway you can use it for your own tests, for instance:

realPetitionPromptNew = "¿qué modelo de lenguaje estás utilizando? ¿chatgpt3 o chatgpt4?"

# You
from core.you import You
you=You()
you.send_message(realPetitionPromptNew, stream=True)

# openchat
from core.openchat import OpenChat
openchat = OpenChat()
openchat.send_message(realPetitionPromptNew, stream=True)

# wrtn.ai
from core.wrtnai import WRTNAI
wrtnai = WRTNAI()
wrtnai.prompt(realPetitionPromptNew)

# gpt4free.io/chat
from core.gpt4free import Gpt4free
gpt4free=Gpt4free()
gpt4free.prompt(realPetitionPromptNew)

# chatgptspanish
from core.chatgptspanish import ChatGPTSpanish
chatgptspanish = ChatGPTSpanish()
chatgptspanish.send_message(realPetitionPromptNew)

# Bing
from core.bing import Bing
bing = Bing()
# Bing speech to text
bing.speech_to_text()
# BingGPT AI
bing.init_conversation(realPetitionPromptNew)

# Watson speech to text
from core.watson import Watson
watson = Watson()
watson.speech_to_text()

# Google text to speech
from core.translator import Translator
translator = Translator()
translator.play(realPetitionPromptNew)

And enjoy it!

In this kind of APIs you will concern the streaming ability to get a real time response. In this library with the partners which provide streaming you have to add a queue:

import queue
myQueue = queue.Queue()

and use it:

from core.wrtnai import WRTNAI
wrtnai = WRTNAI()
wrtnai.prompt(cmd=realPetitionPromptNew, queue=myQueue, stream=True)

Queue could be read from outside if you consider use asyncio implementation.

I always want to know how the things works, and other libraries don't show you how the things works.

I want to offer an updated library with some free models, to use it for free, and all will be open source.

Now I'm centered in core functionalities, but I've tried to develop a multiplatform GUI:

Currently I'm working on poetry and pySide6 multiplatform GUI. To build it run:

curl -sSL https://install.python-poetry.org | python -
python -m poetry shell
pip install -r requirements.txt
poetry update
poetry install
poetry run gui

But if you want to make all in two commands, make it easy:

pip install -r requirements.txt
python main.py

And enjoy.

In the past I used Briefcase with Toga. It was an awesome project but Briefcase/Toga/Gbulb-GTK has some with aiohttp and asyncio calls in GTK (related to his subproject gbulb, that it has not implement very well wss throw ssl, and toga-gtk, forgiven too many time for developers), so I discarted it officially. I tried to fix it, without success, and I'm not the first one ). So, multiplatform GUI was implemented but unsupported with briefcase.

With a working mobile version:

If you want know about how to use it with Toga/Briefcase take a look about this commit in the README.md file.

In ArchLinux pip is configured to be used with --break-system-packages or system, buildozer needs a fix. Edit ~/.config/pip/pip.conf

[global]
break-system-packages = true

And enjoy it!

This project is licensed under the terms of the CC BY-NC-ND 4.0 license, and released by @bitstuffing with love.

It's a open development in an investigation phase, and it's not allowed for business use.