Inspiration:

While exploring the websites of large technology companies, we noticed a common problem: despite having vast amounts of information, many websites struggled to help users quickly find accurate answers from their own content. Users often had to navigate through multiple pages, FAQs, documentation, support articles, and job descriptions just to locate the information they needed.

We also discovered another challenge: many websites either do not have a chatbot at all, or their chatbot lacks the intelligence needed to understand and answer questions based on the website's actual content. As a result, users are often left searching through menus, documentation, support pages, and search results to find simple answers.

Even basic tasks can become frustrating. When visiting a website for the first time, users may spend valuable time searching for things like the login page, pricing information, documentation, API references, support resources, or career opportunities. Even on well-designed platforms, finding the right page can require multiple clicks and unnecessary exploration.

This challenge became even more apparent for students searching for jobs and internships. As students, whenever we found an interesting role on a company website, our next step was usually to open ChatGPT, Gemini, or another AI tool, copy the job description, upload or paste our resume, and ask whether we were a good fit for the role. We repeated this process for every application.

Job hunting today often feels like an endless cycle of searching, comparing, and switching between platforms. Students frequently move back and forth between job portals, company websites, resumes, and AI tools just to understand whether they are a good fit for a role. Instead of focusing on preparing for interviews, improving their skills, and applying with confidence, they spend valuable time gathering information and manually comparing requirements.

At the same time, we realized that websites already contain the information users need, but lack an intelligent way to help people discover, understand, navigate, and act on it.

We asked ourselves:

"What if every website could explain itself, help users instantly find what they're looking for, and provide personalized guidance without leaving the website?"

That idea became Aletheia.

What it does:

Aletheia is an AI-powered website intelligence assistant that transforms any website into an intelligent, conversational experience.

Instead of forcing users to manually search through documentation, support articles, product pages, FAQs, or job postings, Aletheia allows them to interact with a website through natural language and receive accurate, context-aware answers grounded in the website's content.

Users can simply ask questions such as:

  • "Where can I log in?"
  • "Show me the pricing page."
  • "How do I get started with the API?"
  • "What internships are available?"
  • "Am I a good fit for this role?"

Aletheia goes beyond traditional website chatbots by helping users not only find information but also understand and act on it.

It can:

  • Answer questions using website-specific context
  • Guide users directly to relevant pages such as login, pricing, documentation, support, or careers
  • Summarize complex content and documentation
  • Explain processes and concepts in a simple, conversational manner
  • Generate flowcharts and roadmaps from website information
  • Compare resumes against job descriptions found on company websites
  • Identify matching skills, missing requirements, and improvement opportunities
  • Provide personalized insights for students, job seekers, customers, and website visitors

For students and job seekers, Aletheia brings everything together in one place. Instead of copying job descriptions into external AI tools and manually comparing them with a resume, users can simply ask Aletheia. The assistant instantly analyzes the role, compares it against the user's skills and experience, identifies strengths and skill gaps, and provides personalized guidance.

What once required multiple tabs, tools, and conversations can now be accomplished through a single intelligent interaction.

By combining website understanding, intelligent navigation, visual knowledge generation, and AI-powered career assistance, Aletheia helps people spend less time searching and more time taking action.

How we built it:

To bring this vision to life, we built Aletheia as a browser-based AI assistant powered by Retrieval-Augmented Generation (RAG), semantic search, and Large Language Models.

The system operates through a Chrome extension that integrates directly into websites, allowing users to access intelligent assistance without leaving the page. When a user visits a website, Aletheia analyzes the available content, creates a structured knowledge representation, and enables conversational interactions grounded in that context.

Our architecture consists of:

  • A Chrome Extension frontend that injects an AI assistant directly into websites
  • A Python FastAPI backend for orchestration and AI processing
  • Website crawling and content extraction pipelines
  • Document processing and intelligent chunking mechanisms
  • Semantic embeddings for contextual understanding
  • ChromaDB for storing and retrieving website knowledge
  • Retrieval-Augmented Generation (RAG) pipelines
  • Gemini and other LLM providers for reasoning and response generation
  • Structured JSON pipelines for flowchart and roadmap generation
  • Resume-to-job matching capabilities for career assistance

This architecture enables Aletheia to understand website content, retrieve relevant information, generate visual insights, and deliver personalized responses in real time.

Challenges we ran into:

One of the biggest challenges was ensuring that Aletheia responded using the website's actual content rather than generating generic AI answers. Users need trustworthy information, especially when interacting with documentation, policies, support resources, or job opportunities.

Another major challenge involved handling the diversity of websites. Every website has a different structure, navigation system, content organization, and level of accessibility. Extracting meaningful information while maintaining response quality required robust content processing and retrieval strategies.

We also faced challenges in:

  • Reducing hallucinations and improving answer reliability
  • Processing large and complex webpages efficiently
  • Maintaining conversational context across interactions
  • Generating meaningful flowcharts and roadmaps from unstructured content
  • Creating accurate resume-to-job comparisons beyond simple keyword matching
  • Mapping user intent to website navigation paths
  • Integrating browser experiences with backend AI services while maintaining responsiveness and usability

These challenges pushed us to build stronger validation, retrieval, ranking, and context-management systems throughout the platform.

Accomplishments that we're proud of:

We are proud of building a solution that addresses real-world frustrations experienced by millions of website users, students, and job seekers.

Unlike traditional chatbots that provide generic assistance, Aletheia understands the context of the website itself and uses that understanding to deliver relevant, actionable guidance.

Some accomplishments we are particularly proud of include:

  • Transforming websites into intelligent conversational experiences
  • Enabling users to find information without navigating through multiple pages
  • Helping users instantly locate important resources such as login pages, documentation, pricing information, and support resources
  • Assisting students in evaluating job opportunities directly on company websites
  • Building resume-to-job matching capabilities within the browsing experience
  • Generating visual knowledge through flowcharts and roadmaps
  • Creating an end-to-end system that combines website understanding, retrieval, reasoning, navigation assistance, and visualization

Most importantly, we built a tool that helps users move from searching for information to understanding and acting on it.

What we learned:

Building Aletheia taught us that access to information is no longer the primary challenge—understanding information is.

We learned that users often know what they want but struggle to locate, interpret, and connect information scattered across websites. We also discovered that even powerful AI models are only as effective as the context they receive.

Through this project, we gained valuable experience in:

  • Retrieval-Augmented Generation (RAG)
  • Semantic search and vector databases
  • Chrome extension development
  • Large Language Model integration
  • Context-aware AI systems
  • Website content processing and retrieval
  • Diagram generation and knowledge visualization
  • Resume analysis and job matching workflows
  • Intelligent website navigation and information discovery

Most importantly, we learned that AI becomes significantly more useful when it is embedded directly into a user's workflow rather than requiring them to switch between multiple tools and platforms.

Aletheia reinforced our belief that the future of AI is not simply answering questions—it is helping people understand information, make decisions, and take action exactly where they need it.

What's next for Aletheia:

Our vision is simple: make every website self-explanatory, intelligent, and actionable.

Today, Aletheia helps users understand websites through conversational AI, intelligent navigation, visual knowledge generation, and career assistance. However, we believe this is only the beginning.

In the next phase, we plan to expand Aletheia from a website assistant into a comprehensive Website Intelligence Platform that can understand entire websites, not just individual pages.

Our future roadmap includes:

Full Website Understanding

Enable Aletheia to build a complete understanding of multi-page websites, documentation portals, knowledge bases, and enterprise platforms, allowing users to ask questions across an entire website instead of a single page.

Advanced Navigation Assistance

Allow users to instantly locate actions and resources such as login pages, pricing plans, APIs, support articles, forms, and onboarding guides through natural language conversations.

AI-Powered Career Assistant

Enhance resume-to-job matching with personalized career guidance, interview preparation, skill-gap analysis, learning recommendations, and application optimization to help students and job seekers make better career decisions.

Intelligent Knowledge Visualization

Expand our flowchart and roadmap generation capabilities to include architecture diagrams, process maps, learning paths, and interactive visual explanations generated directly from website content.

Agentic Website Interactions

Move beyond answering questions by enabling Aletheia to assist users in completing tasks. Instead of simply explaining how to perform an action, Aletheia will guide users through workflows and help them accomplish their goals more efficiently.

Enterprise Knowledge Integration

Enable organizations to connect internal documentation, knowledge bases, support portals, and business systems, creating a unified AI assistant capable of understanding company-wide information.

Multilingual and Accessibility Support

Make website intelligence accessible to a global audience by supporting multiple languages and providing more inclusive experiences for users with diverse accessibility needs.

Universal Website Intelligence Layer

Our long-term vision is to create a universal intelligence layer for the web—one that can be embedded into any website and instantly transform static information into an interactive, personalized, and intelligent experience.

We envision a future where users no longer have to search endlessly through pages, documentation, and support articles. Instead, every website will be able to understand its own content, answer questions, guide users, generate insights, and help people take action.

Aletheia is our step toward that future.

Built With

Share this project:

Updates