LinkedIn GPT 💙🤖

Warning I am using python3.10

Become the next guru on LinkedIn thanks to ChatGPT and Langchain.

So far it works only on LinkedIn but I hope I made the APIs easy to extend. I recommend to check out the example code for the linkedin ai guru

Warning 🚧 This is a work in progress project

💻 install

git clone git@github.com:FrancescoSaverioZuppichini/LinkedInGPT.git
cd LinkedInGPT
pip install -r requirements.txt

Env Variables

Be sure to export the following variables, you can place them in a .env file

# .env file
LINKEDIN_TOKEN=<YOUR_LINKEDIN_TOKEN>
OPENAI_API_KEY=<YOUR_OPENAI_API_KEY>

To obtain a LINKEDIN_TOKEN check out linkedin-python

Be a Guru

So far, we have the following gurus

LinkedIn AI giga chad 🗿

Run

python gurus/linkedin_ai.py

This will create an automatic post with one of the trending papers from papers with code. This is one example output

And this is the generated post

How it works?

In a nutshell a Guru needs to be able to do

get the content
store the content somewhere
decide which content to post
generate the post
post it

The code for the LinkedIn guru looks like

from langchain.chat_models import ChatOpenAI


from langchain.chains import LLMChain
from langchain.prompts import PromptTemplate

from src.guru import Guru
from src.actions.linkedIn import PostOnLinkedInAction
from src.storages import SQLiteStorage
from src.content_providers import TrendingAIPapersProvider
from src.confirmations.input_confirmation import input_confirmation

prompt = PromptTemplate(
    input_variables=["content", "bot_name"],
    template=Path("prompts/guru.prompt").read_text(),
)

llm = ChatOpenAI(temperature=0)

chain = LLMChain(llm=llm, prompt=prompt)

guru = Guru(
    name="Leonardo",
    content_provider=TrendingAIPapersProvider(),
    storage=SQLiteStorage(),
    action=PostOnLinkedInAction(),
    confirmation=input_confirmation,
    llm_chain=chain
)

guru.run()

content_provider is in charge to get the content we will store and use in the prompt, in this case we get the trending ai papers from paperswithcode
storage stores our contents and retrieves the ones we haven't created yet
action is what we want to do with the content, in our case it will post on LinkedIn
confirmation is a function that will ask for user confirmation, in this case we will just ask the user in the terminal
llm_chain is our langchain chain