Hey reader, and welcome to this introductory blog post where I share the fascinating journey of developing pBot, also known as Pangea Bot. This venture was inspired by my quest for comprehensive information about Pangea, a relatively new player in the market.
Disappointed with the search results, I embarked on creating a solution that would not only provide the information I sought but also serve as a coding bot for the Pangea Community and its array of products.
Motivation and Research
My other initial motivation stemmed from the challenges I faced while trying to integrate APIs and SDKs into my mPoint Hub, a cybersecurity hub based in Uganda. Despite exploring bots like ChatGPT, Bard, and Bing, only Bard yielded information, albeit after tweaking the prompts.
The idea of developing a chatbot utilizing Google Bard Palm API to establish a Pangea context and subsequently training it with substantial Pangea-related data took shape.
Here is ChatGPT's output
Here is Rix's output after context giving
What is pBot?
pBot is a chatbot developed with the primary goal of serving the Pangea community and its ecosystem of products. It also offers extra intel on Domain, IP and user data breach information and awareness alongside providing a secure cyber chatbot system.
About Pangea Cloud
Pangea is a cloud-based security platform that provides a comprehensive suite of security services to help organizations protect their data and systems. Get started here. This project is an inspiration from the hackathon hosted on Hashnode, a free developer blogging platform I have personally used for some good time now.
pBot Components
The development of pBot involved structuring it into three distinct components.
The first, named "Base," is a React app managing authentication and UI based on repository examples.
The second, "Intel," comprises three APIs—User Intel, Domain Intel, and IP Intel—built using Python Flask.
The third component is the actual bot, pBot, constructed on top of Google Palm AI API 2 (Generative Text) using Python Flask. I endowed it with Pangea context for personalized outputs.
pBot Architecture and Workflow
pBot is a sophisticated chatbot designed to serve the Pangea community and its ecosystem. Here's a brief overview of its architecture from ideation to actual implementation:
Authentication:
Handled by the "Base" component using Pangea AuthN.
Manages user authentication and UI presentation.
Intel Component:
Consists of APIs providing intelligence on users, domains, and IPs.
Enhances the chatbot's knowledge base.
pBot Core:
Developed on top of Google Palm API 2.
Learns and customizes responses based on the Pangea context.
Challenges
The integration of the three modules to ensure a seamless user experience across platforms has posed ongoing challenges. Presently, responsiveness on mobile devices remains unavailable.
Additionally, I encountered SDK integration issues with the Python client, leading me to utilize hosted pages. Despite these hurdles, I am optimistic that pBot will be instrumental in overcoming these challenges, given my commitment to being part of the Pangea community.
Here is pBot before Pangea contextualisation
And Here is pBot after Pangea Contextualisation
TRY 🚀
Video Showcase: TLDR
For a visual walkthrough of how pBot operates, I have prepared a comprehensive video demonstration, accessible here.
Next Steps
Continual improvement is the driving force behind pBot's evolution. The goal is to maximize its capabilities and train it with custom Pangea data. This project is open for contributions, available freely, and licensed under MIT.