Transforming PDFs into Conversational Knowledge
Developed by CodeMyPixel for Modezilla LLC (USA), this AI-powered platform turns static documents into dynamic, searchable knowledge bases accessible via chat, SMS, and QR codes.
Project Impact & Results
The platform met critical success criteria for performance, accuracy, and user experience, delivering a robust and scalable solution for automated knowledge retrieval.
The Challenge and The Solution
The Challenge
- Critical knowledge was locked away in unsearchable PDFs, especially low-quality scans.
- Businesses needed a frictionless way for customers to get answers without installing apps or logging in.
- The system had to be fast, secure, and cost-effective, even with millions of documents.
Our Solution by CodeMyPixel
- An end-to-end pipeline: Upload → OCR → Chunk → Embed → RAG Chat with citations.
- A dual-product experience: TextGPT for personal use and IQR.Codes for business with QR/SMS integration.
- A scalable, secure architecture using Supabase with RLS, pgvector, and serverless functions on Vercel.
Powerhouse Technology Stack
Core Platform Features
IQR.Codes for Business
Teams upload documents to create a secure knowledge base. They can then generate QR codes that link customers directly to a branded SMS or web chatbot for instant, automated support.
QR Code Management
Each product has a dedicated management interface to view, edit, and download QR codes. This allows for easy deployment in physical locations, on packaging, or in marketing materials.
Unified Chat Inbox & Logs
All customer interactions, whether from web chat or SMS, are collected in a centralized inbox. Detailed message logs allow for monitoring, analysis, and quality assurance.
Subscription & Billing
The platform integrates Stripe for seamless subscription management. Tiered plans offer different levels of access to features like web chat, SMS channels, and usage limits, complete with a customer billing portal.
Technical Challenges & Solutions
Extracting Text from Image-Based PDFs
Solution: CodeMyPixel implemented an OCR pre-processing step and heuristic cleanup before LLM parsing. This, combined with retries and quality gates, dramatically improved accuracy on difficult, low-quality scans.
Sub-Second Retrieval at Scale
Solution: We tuned pgvector indexes, implemented efficient batch writes, and used cosine similarity with filtered scopes. Lightweight metadata caching was added for frequently accessed queries to ensure speed.
Strict Multi-Tenant Isolation
Solution: A robust, RLS-enforced schema was designed with service-role segregation. Signed URLs and scoped policies for Supabase Storage ensure that organizational data remains strictly separated and secure.
Frequently Asked Questions
Who developed the TextGPT + IQR.Codes platform?
The platform was proudly designed, developed, and managed end-to-end by the expert team at CodeMyPixel for our client, Modezilla LLC (USA).
What is the core function of the platform?
It transforms uploaded PDFs—even difficult, image-based scans—into a searchable knowledge base. Users and customers can then ask questions and get instant, accurate answers through a conversational chat interface, with responses grounded in the source documents.
How does IQR.Codes help businesses?
IQR.Codes allows businesses to provide instant, automated customer support. They can upload product manuals, FAQs, or policies, and generate a QR code. When a customer scans the code, they can immediately start a conversation via SMS or web chat to get answers, reducing support load and improving customer satisfaction.
