Built by
Simon Chow

A voice-first AI web app built for hands-free interaction. Users call in, speak naturally, and get real-time AI responses with zero latency. Supports extended conversations without needing to tap a single button. Built for multitasking scenarios: driving, cooking, working out, or when typing isn't an option. Features WebRTC-based real-time audio streaming with DeepSeek reasoning, ElevenLabs voice synthesis, and whisper-level speech recognition for Cantonese and English.

Snap a photo of any business card and let AI extract, organize, and sync the contacts instantly. Uses GPT-4 Vision for OCR with multi-language support (Chinese, English, Japanese), auto-detects fields like email, phone, company, and title. Features a smart contact feed, CSV export, and CRM integration. Handles non-standard card layouts, low-light captures, and tilted angles gracefully. Designed for sales professionals and networkers who collect dozens of cards at events.

An intelligent photo portfolio that uses AI vision to automatically curate, categorize, and present your best visual work. Upload photos and the system auto-tags them by subject (portrait, landscape, product, event) using CLIP embeddings. Features smart cropping with face detection, dynamic masonry layouts, and intelligent arrangement prioritizing highest-rated images. Built for photographers, designers, and creative professionals who want an AI-assisted portfolio with zero manual tagging.

A streamlined Stripe payment flow with beautiful UX. Handles subscriptions, one-time payments, and complex checkout logic including promo codes, trial periods, and invoice management. Features real-time payment status updates, smooth error recovery, and responsive mobile checkout. Currently resolving a CORS configuration issue with the Netlify deployment before relaunch. Built for SaaS products needing a clean, white-label payment experience.

A full-featured project management platform built for AI-assisted workflows. Features kanban boards with drag-and-drop, task CRUD with granular status tracking, team collaboration with real-time updates, and smart dashboards with completion metrics. Backed by AppWrite on Singapore infrastructure with a Next.js API proxy layer for edge performance. Designed to be operated directly by AI agents — tasks can be created, assigned, and tracked via Telegram bot integration without human intervention.

Real-time dashboard for monitoring API consumption and spending across multiple AI providers (OpenAI, DeepSeek, ElevenLabs, KIE, PiAPI, Replicate). Features interactive Chart.js visualizations with daily/weekly/monthly views, cost breakdowns by provider and model, and automated daily Telegram reports. Data is stored in a local JSON records file and auto-deployed via launchd cron. Built to solve the problem of runaway AI API costs — gives you visibility into exactly what you're spending and on which models.

An autonomous phone agent deployed on an Oppo R15 Android device via ADB. Runs AutoGLM to automate on-device tasks: opening apps, filling forms, reading notifications, and interacting with Chinese app ecosystems (WeChat, Alipay, Douyin). Uses Python-based ADB automation scripts with AI-powered visual understanding to navigate the UI. Still in active beta development — working on reliability and complex multi-step task chains. Open-source framework being customized for Hong Kong use cases.

An autonomous voice agent that answers calls, transcribes speech in real time, and responds with AI-generated answers. Integrates with Twilio for phone routing and DeepSeek for reasoning. Handles appointment booking and FAQ flows for call centers and small businesses. Supports Cantonese, Mandarin, and English speech recognition via Whisper, with ElevenLabs voice synthesis for natural-sounding responses. Built for small businesses that need a 24/7 AI receptionist without the overhead of a human operator.