Introduction
Forget the Past—Bb's Got Your Back
How many times have you seen something online and couldn't find it again? That product you were researching, that article you read last week, that tutorial video you watched—all lost in the endless sea of browser history.
Bb (Boring Brains) solves this by intelligently recording your browsing activity and making it searchable using natural language. Ask "What was that blue sneaker I saw last week?" and Bb will find it for you—all while keeping your data completely private and self-hosted.
Built in 10 Days - Privacy-First Design
Bb brings the power of Windows Recall to your browser—but with privacy at its core. Everything stays self-hosted, encrypted, and under your complete control.
Key Features
AI-Powered Search
Query your browsing history using natural language. Ask anything and Bb finds it in seconds.
Privacy First
All data processed locally on your device or self-hosted server. No cloud, no tracking, no third parties.
Perfect for Shopping
Find that product, price, or store you saw days ago. Works for prices, links, images—anything visual.
Research & Study
Perfect for students and researchers. Never lose track of sources, articles, or references.
Vision AI
LLaMA 3.2 Vision-powered OCR and image understanding for accurate content retrieval.
Easy Setup
Docker-based deployment. Run locally or on a remote server with simple commands.
The Story Behind It
The idea for Bb originated from my fascination with Microsoft's Windows Recall feature—a tool that created a searchable timeline of activities like apps and websites. While this concept was groundbreaking, it drew intense criticism over privacy concerns, with many comparing it to spyware.
Though I shared the excitement for such a tool, I was equally committed to protecting user privacy. This led me to develop a similar tool that puts user autonomy first. The result is Bb: a Chrome extension that brings powerful recall to your browsing experience while keeping your data secure and self-hosted.
What It Does
Bb transforms your browsing history into a powerful searchable timeline. It captures screenshots of your activities—shopping, studying, researching—and allows you to retrieve information using natural language queries.
Real-world examples:
- ✓ "Where was that red T-shirt from for 499 rupees?"
- ✓ "What was the price of the television I saw on Xiaomi's website?"
- ✓ "Show me that YouTube tutorial I watched last week"
- ✓ "Find the link to that research paper I was reading"
Within seconds, Bb provides the answer along with the relevant image. The best part? Bb is entirely self-hosted, keeping your data under your control. Your data remains exclusively on your device or server and is only processed by open-source models that never train on your personal information.
Our Technical Approach
Bb leverages LLaMA 3.2-90B-Vision-Instruct powered by SambaNova, with the backend hosted locally on your device or server.
Core Components
- • LangChain: Agent-based & prompt chaining
- • LLaMA Vision: Advanced image understanding
- • FAISS: Vector search optimization
- • SambaNova: Open-source AI inference
Infrastructure
- • FastAPI: High-performance backend
- • Docker: Simplified deployment
- • Chrome Extension: User interface
- • Local/Self-Hosted: Complete privacy
💡 The codebase is built for production with scalability and simplicity at its core, allowing users to deploy on their own hardware with just Docker commands.
System Workflow
Bb operates through a sophisticated multi-component architecture that handles authentication, screenshot capture, OCR, embeddings, vector search, and AI inference—all happening locally:
Login & Authentication
Secure one-time token authentication ensures only authorized access. Docker-based login creates a login token during server setup.
Screenshot & Processing
Chrome extension captures screenshots of active tabs every 12 seconds. Images are processed locally using vision AI for OCR and content extraction.
Search & Retrieval
SambaNova creates embeddings, FAISS performs fast vector search, and LangChain agents rank results by relevance. All done every 45 seconds.
📊 Complete System Architecture Diagram
Detailed system architecture showing all components and data flow
Installation on Linux
Bb is designed for easy self-hosting. Here's how to get started:
Prerequisites
# Linux system with:
Docker Compose
ffmpeg, libsm6, libxext6
Step 1: Install Dependencies
sudo apt install docker-compose ffmpeg libsm6 libxext6 -y
Step 2: Clone & Setup
git clone https://github.com/AyushDhimann/BoringBrains
cd BoringBrains/backEnd
cp .env.example .env # Add SAMBANOVA_API_KEY
Step 3: Build & Deploy
docker-compose build
docker-compose up
# Save the one-time login token that appears
Step 4: Configure Chrome Extension
- ✓ Open the Bb Chrome extension
- ✓ Paste the login token
- ✓ Enter your server's IP (localhost for local)
- ✓ Enable screenshot permissions
How to Use Bb
1️⃣ Enable Permissions
Allow screenshot permissions for your Chrome browser to let Bb capture your browsing activity.
2️⃣ Browse Freely
Bb automatically captures screenshots every 12 seconds as you browse. Opt-in control means you're always in charge.
3️⃣ Query Anytime
Type natural language queries into the Bb extension. Search for specific items, content, or anything from your browsing history.
4️⃣ Get Results
Bb retrieves associated screenshots and data within seconds. Results include the image, timestamp, and relevance score.
AI Architecture & Capabilities
Bb leverages advanced AI techniques for production-level performance:
LangChain Agent-Based Architecture
Uses agent-based and prompt chaining techniques to create a multi-step reasoning process. The agent breaks down complex queries into manageable tasks, performs semantic searches, and ranks results intelligently.
Vision-Based OCR
LLaMA 3.2-90B-Vision extracts text and visual content from screenshots. The system understands images contextually, not just keyword-matching, enabling accurate retrieval even for abstract queries.
Vector Embeddings & FAISS Search
SambaNova generates semantic embeddings for all captured content. FAISS provides fast approximate nearest neighbor search, enabling millisecond-level retrieval from thousands of screenshots.
Contextual Ranking
Results are re-ranked using contextual relevance scoring. The system understands nuance, temporal context, and visual similarity to surface the most relevant results first.
Overcoming Technical Hurdles
Privacy vs. Functionality
Ensuring Bb's features were robust while keeping all data processing local and secure. Solved through self-hosted infrastructure and end-to-end encryption.
AI Integration & Performance
Optimizing LangChain's agent-based architecture for fast responses. Balanced between accuracy and speed through caching and efficient embedding search.
User-Friendly Self-Hosting
Simplifying Docker deployment for users unfamiliar with server management. Automated setup scripts and clear documentation made the process accessible.
Key Accomplishments
⏱️ Built in Just 10 Days
Completed a full-stack, production-ready application with complex AI integration in minimal time—a testament to focused execution and architectural planning.
🔒 Privacy-First Architecture
Successfully built a tool that respects user privacy by keeping all data self-hosted, encrypted, and processed locally—no cloud storage, no data training, no third parties.
⚡ Fast AI-Powered Recall
Seamlessly integrated advanced vision AI to deliver fast and accurate retrieval from thousands of screenshots within seconds.
📦 Accessible Deployment
Designed straightforward Docker setup with clear documentation, making self-hosting accessible to a wide range of users regardless of technical background.
Lessons & Insights
Building Bb taught me invaluable lessons about building AI systems while respecting privacy:
🏗️ Scalable AI Systems
Learned how to architect LangChain systems with proper caching, async processing, and vector databases for production-scale performance.
🔐 Privacy by Design
Understood how to balance powerful features with strong privacy guarantees—showing that advanced AI and user autonomy aren't mutually exclusive.
🖼️ Vision AI Integration
Mastered LLaMA Vision for OCR, contextual understanding, and intelligent content processing—critical for real-world multimodal applications.
📦 Docker & Deployment
Learned how to containerize complex AI applications for easy, reproducible deployment—enabling users to run sophisticated systems locally.
The Roadmap Ahead
Bb was designed to be cross-platform from the start. Future enhancements include:
📱 Multi-Platform Support
Extend beyond Chrome to Safari, Firefox, and Edge. Bring the same privacy-first recall to all browsers.
🔑 Biometric Auth
Add fingerprint and facial recognition for device-level access control.
⚙️ Customizable Retention
Allow users to set their own parameters for data capture frequency and storage duration.
🌍 Open Source
Further open-source the project to enable community contributions and improvements.
Technology Stack
AI & ML
Backend & Infrastructure
Frontend
Other Tools
Watch the Demo
Get Started
✨ Ready to Remember Better?
Download Bb today and start building your searchable digital memory. Privacy, speed, and intelligence—all under your control.
"Forget the past—Bb's got your back, always and forever!"