Mobile Header with Scroll Control
Contents

Project Introduction

The Customised ChatBot project aims to develop an intelligent chatbot capable of providing detailed answers to user queries. This chatbot extracts and embeds data from various URLs, including video and SVG content. It references these sources while answering user queries, offering comprehensive responses and citing sources. The chatbot leverages cutting-edge technologies to ensure accurate and relevant information delivery.

HLD

Data Extraction

URLs, including video and SVG content, are crawled and embedded.

Database Layer

MicrosoftSQL stores chat history, embedding details, and other relevant data.

Embedding Storage

Pinecone is used to store the embeddings.

Backend Layer

FastAPI handles API requests and responses, ensuring efficient communication.

AI Integration

OpenAI and Langchain process user queries, generate responses, and reference relevant sources.

Discovery Phase Details and Process

The discovery phase was critical for understanding the project requirements and defining the implementation strategy. Key activities included:
Requirement Gathering

Engaged with stakeholders to gather detailed requirements and user expectations.

Feasibility Study

Analyzed technical feasibility and identified potential challenges.

Technology Selection

Selected MicrosoftSQL, FastAPI, OpenAI, Langchain, and Pinecone as the tech stack for their robustness and compatibility.

Data Collection Strategy

Designed a strategy for extracting and embedding data from various URLs, including video and SVG content.

Prototyping

Created initial prototypes to validate the approach and gather early feedback.

Libraries Used

The project utilizes several libraries to enhance functionality:

SQLAlchemy: For database operations and ORM.

BeautifulSoup & Requests: For web scraping and data extraction from URLs.

Langchain: To manage and process natural language queries.

Pinecone: For storing and managing embeddings.

Databases Used

Microsoft SQL: The primary database for storing chat history, embedding details, and other related data.

Pinecone: Used for storing embeddings of extracted content, enabling quick and efficient retrieval during chatbot interactions.

Integrations Performed

Several integrations were essential to achieve the project’s objectives:

OpenAI API

Integrated for natural language processing to understand and interpret user queries.

Langchain

Used for advanced language processing and chaining multiple language models.

Pinecone

Integrated to store and manage embeddings of the extracted content.

FastAPI

Ensures smooth communication between the backend and the AI models, providing a responsive user experience.

Data Extraction Tools

Integrated libraries for scraping and extracting data from various URLs, including video and SVG content.

The latest from our den

View the blog
  
Staff Augmentation

We provide highly skilled, on-demand technical...

Artificial Intelligence

Harness AI to optimally extract value from your...

Cyber Security

At Ziel Global, we provide cutting-edge...

Cloud Engineering

At Ziel Global, we empower businesses to harness...

Mobile App Development

Outsourcing your mobile app development will...

Custom Software Development

Our Offshore Software Application Development...

UI/UX

At Ziel Global, we understand that exceptional...

ERP Consultancy & Delivery

With a global presence, we specialize in cutting-edge ERP solutions, leveraging

to optimize operations and unlock business potential.