Video Showcase

Demo videos of prototypes & products, each with a full explanation

Demo Videos

Walkthroughs of what we build - each clip comes with a complete explanation of the problem it solves and how it works.

AIoT 0:16 June 2026

Handheld Sub-GHz RF Multi-Tool - Custom Embedded Spectrum Analyzer

A live demo of a custom-built handheld RF multi-tool engineered in-house for sub-GHz diagnostics and embedded development. The device pairs an ESP32 microcontroller with a CC1101 sub-GHz transceiver and an SMA antenna, all housed in a compact 3D-printed case with a colour TFT display and a carousel UI. From the RF menu it offers a real-time spectrum analyser, Scan/Copy, Simulate, Custom Sub-GHz and SquareWave tools. In the clip the spectrum view is centred on the 433.92 MHz ISM band (432.9-434.9 MHz) and the live waveform reacts as a 433 MHz remote is exercised nearby - letting an engineer see exactly which frequency a device transmits on and how clean its signal is. It's a practical bench instrument for testing remotes, sensors and IoT radios, and a showcase of our hardware capability: custom PCBs, RF front-ends, firmware and enclosure design delivered as one integrated product.
AI Vision 0:35 December 2025

AI Defect Detection - Industrial Anomaly Inspection with Heatmaps

A live demo of an AI Defect Detection system for industrial quality control. It learns what a good part looks like from a bank of reference image patches (a ResNet feature extractor with a coreset memory bank), then inspects each test image and flags anomalies it has never seen before - no defect examples needed for training. For every inspection it returns a clear PASS or DEFECT DETECTED verdict with an anomaly score against a decision threshold, confidence and peak-intensity readings, the number of affected patches, and a red heatmap overlay that pinpoints exactly where the flaw is on the part. Detection sensitivity and sampling are tunable for the speed-vs-accuracy trade-off (industrial QC typically runs 0.6-0.8). In the clip it inspects precision circular metal components and localises surface defects in seconds - the kind of automated visual inspection that replaces slow, inconsistent manual checking on a production line.
AI Agent 1:24 April 2025

VoidAgent - Autonomous AI Browser Agent with Live Task Streaming

A live demo of VoidAgent, an autonomous AI browser agent that completes real web tasks on its own. Given a goal, it drives a real Chrome browser - navigating sites, searching, and clicking through pages - while streaming every step it takes in real time: each action, the element it clicked, timestamps, and a final DONE. In the clip it works on Tokopedia, an Indonesian marketplace, carrying out the instruction 'find the product esp32-c3 and add two of them to the cart' - searching, opening listings, and adding the items to the cart entirely on its own. A Start Task / Stop Task control runs the agent, and the streaming log makes its reasoning and actions fully transparent. It's a practical look at agentic automation - handing routine web work like research, shopping, data entry, and repetitive flows to an AI that operates the browser exactly like a person would.
AI Vision 1:16 January 2026

Retail Store Analytics - AI Shopper Tracking, Shelf Zones & Dwell Time

A live demo of a Retail Store Analytics system that turns an ordinary store camera into shopper-behavior insight. Using AI pose detection, it tracks every shopper in the aisle as a skeleton, assigns each a persistent ID, and watches custom ROI zones drawn in front of the shelves - here a zone labelled 'Rak Makanan Kucing' (Cat Food Shelf). For each zone it counts how many people are in it right now, total entries, and the average dwell time, and logs per-person dwell (e.g. Person 42 - 3s) along with total shelf interactions. The result is a real-time picture of which displays draw attention and how long shoppers actually linger - data retailers can use to optimise shelf placement, promotions, and store layout, all from a single camera feed.
AIoT 0:22 June 2025

Edge AI Vision - Blister-Strip Counting with On-Device Inference & MQTT

A live demo of an edge AI-vision system that counts stacked pharmaceutical blister strips in real time. A small camera-equipped edge device runs the detection model on-device - the dashboard's live logger streams its per-frame timing (preprocess / inference / postprocess) - and classifies how many strips are stacked, showing the result as 1 / 2 / 3 strips with a confidence score (here 2 strips at 93%). An adjustable confidence threshold tunes sensitivity, and the device reports over WiFi with MQTT for IoT integration. The clip is shown as a merged view - the physical setup on top, the dashboard below - so you can watch the hardware and the AI response at the same time. A practical example of on-device vision for fast, automated counting and quality checks in pharmaceutical packaging and inventory.
AR/VR 0:49 May 2026

Vanguard - Augmented Reality First-Person Shooter Game

A live demo of Vanguard, a mobile augmented-reality first-person shooter that drops the game into your real surroundings. Through the phone camera, virtual enemies are placed into the actual space - here an outdoor cafe - and the player aims and fires at them in first person, complete with a sweeping radar, health and ammo HUD, kill counter, level, and match timer. It's a full game, not just a tech demo: a branded menu with Single Player, Multiplayer, Store, Global Rank, and Training modes, plus a player profile. Vanguard shows our AR stack at play - real-time camera tracking, 3D enemies anchored into the world, and responsive shooter mechanics running on a phone.
AI Vision 1:52 December 2025

E-TLE Traffic Monitoring - AI License-Plate OCR & Bus Passenger Counting

A live demo of the E-TLE Traffic Monitoring System, an AI computer-vision dashboard that watches two camera feeds at once. On the road/dashcam side it detects lanes and vehicles, marks a detection zone, and reads license plates with OCR (confidence scores shown), automatically logging traffic violations with plate number, vehicle type, and time. On the bus-interior side it counts passengers, tracking people crossing an in/out line and keeping a running total with a timestamped passenger log. Together it turns ordinary camera streams into structured, auditable data for traffic enforcement and public-transport monitoring - vehicle counts, plate reads, violations, and passenger flow, all in real time.
AIoT 2:00 June 2026

Memoir - USB Voice Recorder to AI Journal, in the Browser

A live demo of Memoir, a web app that turns a small USB voice recorder into an AI-written journal - with no app to install. The hardware recorder captures audio to a microSD card; plug it in and Memoir connects to the device directly from the browser, pulls the recordings off it, transcribes the speech, and generates a clean written journal entry from what was said. The dashboard tracks how many recordings are on the device, how many have been transcribed, total captured time, and storage used, with one-tap Connect / Disconnect and a Generate button. In the clip, a real Indonesian voice note is transcribed and turned into a dated journal entry - showing how custom recording hardware, in-browser device access, speech-to-text, and AI text generation come together into one seamless capture-to-journal workflow.
AR/VR 1:00 May 2026

Augmented Reality HUD - AR Overlay with IoT Hardware Integration

A live demo of an augmented-reality experience running entirely on a smartphone. The phone's camera looks at the real workspace and renders an interactive, game-style HUD on top of it in real time - a sweeping radar minimap, health and status indicators, and 3D elements anchored into the live scene. On the desk sits a custom-built edge device, a wired microcontroller board that feeds into the experience, showing how AR graphics and physical IoT hardware can work together. The result is a responsive, low-latency mixed-reality layer that turns an ordinary camera view into an immersive, data-rich interface - the kind of AR/hardware integration we build for product demos, training, and interactive installations.
AI Vision 0:56 November 2025

Nusatec Vision - AI Expiry-Date Inspection & Counting

A live demo of Nusatec Vision, a computer-vision inspection system that checks products for a printed expiry date and counts them in real time. An edge device runs the detection on-site and streams results to a dashboard that splits the tally into items that have an expiry date and items that don't (ADA / TIDAK ADA EXP DATE), grouped by class. From the panel an operator can manage class labels, build reference cards, talk to the hardware over a serial link, and control the run - execute or pause, with adjustable speed - turning manual date-checking into an automated, auditable inspection line.
IoT 1:04 January 2026

Nusamesh - Off-Grid LoRa Mesh Messaging

A live demo of Nusamesh, our off-grid LoRa mesh communication platform. Several low-power mesh nodes automatically form a self-healing network, and a paired phone app sends and receives text messages across them with no internet or cellular connection at all. In the clip, messages are typed on the handset and relayed node-to-node in real time - decentralized communication that keeps working where conventional networks can't reach, built for remote sites, maritime operations, and disaster response.