ConvoCatalyst : LinkedIn Insights & Ice-breakers

A LangChain powered LLM model that provides you a summary, interesting facts, topics of interest and an ice-breaker to help start a conversation with a person by scraping their LinkedIn profile (and Twitter profile ?).

Working:

Getting Results for Harrison Chase, founder & CEO - Langchain.

Input:

name = "Harrison Chase"

Note: If the model scrapes the profile of another person with same name rather than the person you are looking for : Try providing a bit of unique detail into the name. For eg:

if "Harrison Chase" does not get you desired results , try something like "Harrison Chase LangChain"

Scraping and Prompt:

Given name of person, use lookup agent with the help of the tool get_profile_url() (SerpAPI) to find out the person's linkedin_profile_url. All this powered by our llm model.

LLM Model being used is gpt-3.5-turbo from the OPENAI API.

Passing the linkedin_profile_url as the input down here to the {information} placeholder as prompt template to our langchain powered llm model.

    summary_template = """
        given the Linkedin information {information} about a person, i want you to create:
        1. a short summary
        2. an interesting fact about them
        3. A potential topic of interest. Keep it short.
        4. 1 creative ice-breaker to open a conversation with them.
        """

Result:

The most interesting and mindblowing thing here is the demonstration of the thought process of the agent which basically acts as a brain to the model for the steps it needs to take to reach desired output.

Also, the results seemed pretty convincing! LangChain is cool!

Run the Project

Environment Variables

To run this project, you will need to add the following environment variables to your .env file

PYTHONPATH=/{YOUR_PATH_TO_PROJECT}/ice_breaker

OPENAI_API_KEY

PROXYCURL_API_KEY

SERPAPI_API_KEY

TWITTER_API_KEY

TWITTER_API_SECRET

TWITTER_ACCESS_TOKEN

TWITTER_ACCESS_SECRET

[twitter scraping is not implemented now due to issues with twitter api, not allowing certain api calls in some countries/regions]

Run Locally

Clone the project

  git clone https://github.com/SARIT42/convo-catalyst.git

Go to the project directory

  cd ice_breaker

Use a pipenv environment.

Install dependencies

  pipenv install

run the icebreaker.py file.

  pipenv run app.py

Twitter Scraping.

Getting the following error:

tweepy.errors.Forbidden: 403 Forbidden When authenticating requests to the Twitter API v2 endpoints, you must use keys and tokens from a Twitter developer App that is attached to a Project. You can create a project via the developer portal.

which to my knowledge, as i figured out was because of some twitter api policy changes applicable to certain regions/countries. So nothing much can be done on the developer side ig.

This halts the twitter account information scraping but the implementation code for the twitter_lookup_agent and the tweets processing in twitter.py has been provided in the respective sections.

Any help in resolving the above issue is highly appreciated.