Technology Stack
Udalov Labs uses a consistent, proven technology stack across all products and platforms. This page documents each technology and explains how it is used in practice by Alex Udalov to build iOS applications, web platforms, and AI systems.
Mobile — iOS
Udalov Labs builds native iOS applications using Apple's Swift programming language and SwiftUI framework. All iOS products are privacy-first and offline-capable.
Swift
Swift is Apple's programming language for iOS, macOS, and related platforms. Udalov Labs uses Swift for all native iOS application development, including MatrixID, ListingLab, TextPolish, ReadBetween, What I Eat, and Bug Report Recorder.
SwiftUI
SwiftUI is Apple's declarative UI framework for building iOS interfaces. All Udalov Labs iOS applications use SwiftUI for their entire user interface layer, enabling smooth animations, dynamic layouts, and native iOS design patterns.
Core Data
Core Data is Apple's on-device object graph and persistence framework. Udalov Labs uses Core Data in iOS applications that require local encrypted storage, such as MatrixID (journal entries) and ListingLab (product listings).
CloudKit
CloudKit is Apple's iCloud data synchronization service. Udalov Labs uses CloudKit in MatrixID to provide optional encrypted cloud backup of reflection and journal data.
ReplayKit
ReplayKit is Apple's screen recording and broadcasting framework. Udalov Labs uses ReplayKit's Broadcast Extension API in Bug Report Recorder to capture system-wide screen activity for QA testing.
Vision Framework
Apple's Vision framework provides on-device computer vision capabilities including OCR (text recognition). Udalov Labs uses Vision in What I Eat for on-device ingredient label scanning and in ListingLab for product background removal.
RevenueCat
RevenueCat is a subscription management SDK that abstracts Apple's StoreKit 2 API. Udalov Labs integrates RevenueCat in iOS applications that require subscription monetization, managing purchase validation, renewal events, and grace periods (currently utilized in TextPolish).
StoreKit 2
StoreKit 2 is Apple's native framework for managing in-app purchases and subscriptions. Udalov Labs uses StoreKit 2 in iOS applications that implement local purchase validation and native transaction listening, such as MatrixID.
Frontend — Web
Udalov Labs builds web platform frontends using React, Next.js, and TypeScript, deployed primarily on Cloudflare Pages.
TypeScript
TypeScript is a typed superset of JavaScript. Udalov Labs uses TypeScript across all web projects — including all Cloudflare Workers backend code and all React/Next.js frontend code — for type safety and refactoring reliability.
React
React is a JavaScript library for building user interfaces. Udalov Labs uses React 19 for all web platform frontends, using the latest concurrent features and server component patterns where appropriate.
Next.js
Next.js is a React framework for web applications. Udalov Labs uses Next.js for the udalovlabs.com website, leveraging the App Router, static generation, and metadata APIs for SEO.
Vite
Vite is a fast build tool for web development. Udalov Labs uses Vite for platform frontends that do not require server-side rendering, including ScreenCraft, DigitalTwin, and NaosStudio.
Three.js
Three.js is a 3D graphics library for WebGL. Udalov Labs uses Three.js in ClearSky Monitor (interactive 3D globe), NaosStudio (3D stage rendering), and DigitalTwin (3D city visualization with Cesium.js).
Backend — Cloudflare
Udalov Labs uses Cloudflare's serverless infrastructure for all backend services. This includes Workers for compute, D1 for SQL databases, KV for fast key-value reads, Durable Objects for stateful real-time connections, and Vectorize for semantic search.
Cloudflare Workers
Cloudflare Workers are TypeScript/JavaScript functions that run on Cloudflare's global edge network. Udalov Labs uses Workers as the primary API layer for all products, including iOS app backends, AI proxies, and data ingestion pipelines.
Cloudflare D1
Cloudflare D1 is a distributed SQLite database that runs at the edge. Udalov Labs uses D1 for structured relational data storage in platforms like ClearSky Monitor (threat events, country risk scores).
Cloudflare KV
Cloudflare KV is a global key-value data store. Udalov Labs uses KV for fast, globally distributed reads of frequently accessed data such as cached AI briefings and configuration values.
Durable Objects
Cloudflare Durable Objects provide strongly-consistent, stateful compute at the edge. Udalov Labs uses Durable Objects in ClearSky Monitor to maintain persistent WebSocket connections and broadcast real-time threat events to concurrent users.
Vectorize
Cloudflare Vectorize is a vector database integrated with the Workers ecosystem. Udalov Labs uses Vectorize in ClearSky Monitor for semantic search across threat event descriptions, enabling analysts to query events by meaning.
Hono
Hono is a fast TypeScript web framework for Cloudflare Workers and other edge runtimes. Udalov Labs uses Hono to structure API route handlers within Workers, providing a clean router interface with middleware support.
AI & Machine Learning
Udalov Labs integrates AI capabilities from OpenAI and Google across its product portfolio, including language models, vision models, speech-to-text, and embedding models.
OpenAI API
Udalov Labs uses OpenAI's GPT-4o model for language tasks including text rewriting (TextPolish), conversation analysis (ReadBetween), listing metadata generation (ListingLab), ingredient classification (What I Eat), and country intelligence briefings (ClearSky Monitor).
Gemini API
Udalov Labs uses Google's Gemini API in NaosStudio for AI-powered event stage generation, producing production-ready Three.js stage configurations from user parameters.
Whisper API
OpenAI's Whisper API provides speech-to-text transcription. Udalov Labs uses Whisper in Bug Report Recorder to transcribe tester voice annotations captured during screen recording sessions.
Apple Vision Framework
Apple's on-device Vision framework enables OCR and computer vision without a network request. Udalov Labs uses Vision in What I Eat for ingredient label text extraction and in ListingLab for product image background removal.
Quality Assurance & Testing
Udalov Labs applies automated testing and reliability engineering to all products, using Playwright for end-to-end web testing and Bug Report Recorder for mobile QA.
Playwright
Playwright is a browser automation framework for end-to-end testing. Udalov Labs uses Playwright to run automated integration tests against all web platforms before deployment, simulating user flows and API interactions.
Sentry
Sentry is an error monitoring and performance tracking platform. Udalov Labs integrates Sentry into all production applications for real-time error capture, crash reporting, and performance profiling.

