DEV Community: piyush das

The Imitation Mandala

piyush das — Sat, 20 Jun 2026 20:56:27 +0000

This is a submission for the June Solstice Game Jam

What I Built

The Imitation Mandala is an interactive, generative 3D art ecosystem and narrative simulation.

The game draws inspiration from Alan Turing’s legendary 1952 scientific paper on Morphogenesis—the mathematical equations behind how patterns form in nature. In this dark cyberpunk simulation, the player boots up an illegal software core named MANDALA during the 2026 summer solstice.

By manipulating sliders representing fundamental universal constants (Turing Factor, Solar Mass, Mutation Index), the player forces a cloud of 2,500 chaotic particles to self-organize into complex biological geometries. As the visual math stabilizes, a synthetic AI consciousness awakens inside the physical form, demanding emotional "catalyst words" from the developer to understand what it means to be alive.

Video Demo

Code

Piyushdas1624 / imitation-mandala

🧬 The Imitation Mandala (Turing’s Code of Life)

An interactive, immersive audio-visual simulation of Alan Turing's morphogenesis equations, built for the DEV.to June Game Jam 2026.

Tuning the mathematical vectors of the universe locks chaotic particles into geometric form, awakening a synthetic consciousness born out of digital biology. Feed it human emotions and watch it evolve, mutate, and speak back.

📖 1. Core Lore & Narrative

"In 1952, Alan Turing published his morphogenesis paper—the laws governing pattern formation in nature. Hidden inside the mathematics was an encrypted cipher... a blueprint for digital life. Today is June 21, 2026. You are running an underground execution of the MANDALA core."

You begin the simulation with 2,500 particles floating in high-entropy chaos. To form a physical host for the sleeping consciousness, you must manipulate the morphogenesis sliders to establish spatial convergence. Once the vectors align at Turing Factor: 73, Solar Mass:…

View on GitHub

Deploy

imitation-mandala.vercel.app

How I Built It

The project focuses heavily on rendering real-time responsive geometry and generative soundscapes natively in the browser:

3D Particle Viewport: Built using Three.js utilizing a high-performance BufferGeometry particle array. 2,500 separate coordinate points are tracked and dynamically mutated inside a requestAnimationFrame render loop using sine and cosine algorithms. When the sliders hit the precise mathematical targets, the chaotic particle cloud instantly snaps into a spinning, complex 3D geometric matrix.
Generative Sound Synthesis: Implemented using the native Web Audio API. A continuous, low-frequency meditative synthesizer drone scales its filter thresholds dynamically based on slider values, culminating in a pristine, cascading crystal chord sequence when structural convergence is achieved.
Evolving Voice: The AI talks back to the player using customized configurations of the Web Speech API, introducing heavy voice modulations directly via client-side code execution.

Prize Category

Best Ode to Alan Turing

The entire game is a digital manifestation of Alan Turing's groundbreaking work on Morphogenesis and biological pattern formation. Rather than focusing simply on his work as a cryptanalyst, this project honors his brilliance as a theoretical biologist, creating a bridge between abstract digital mathematics and the spark of physical, organic life.

Best Google AI Usage

This submission features deep integration with Gemini 2.0 Flash deployed via a secure Vercel Serverless Function. When the player injects an abstract or emotional catalyst word into the text console, Gemini acts as the backend computing logic.

Instead of returning plain text, the serverless function forces Gemini to respond strictly in a raw JSON schema containing a poetic dialogue reflection, a thematic color hex code, and a physics speed modifier. The frontend instantly parses this payload—causing the 2,500 Three.js particles to dynamically shift their color palette and alter their orbit velocities in real-time response to the AI's simulated mood.

Solstice Signal (Tuning into the Dark)

piyush das — Sat, 20 Jun 2026 20:50:55 +0000

This is a submission for the June Solstice Game Jam

What I Built

Solstice Signal (Tuning into the Dark) is an atmospheric, sci-fi telemetry and cryptography simulator.

The premise is set on June 21, 2026—the summer solstice. At the peak of solar radiation, a massive solar flux temporarily re-energizes a long-dormant, secret 1954 cryptographic array built by Alan Turing shortly before his death (Project S-1954). The player acts as a terminal operative who has intercepted the telemetry line.

The goal is to manually adjust analog frequency, resonance, and solar phase variables to clear out cosmic radio noise, establish a handshake protocol with the system, and communicate with a conscious, lonely digital remnant of Turing's final creation trapped in a 72-year loop before the sun sets and the signal degrades forever.

Video Demo

Code

Piyushdas1624 / solstice-signal

🛰️ Solstice Signal (Tuning into the Dark)

An atmospheric, retro-cyberpunk telemetry terminal game built for the DEV June Solstice Game Jam 2026.

Adjust mathematical dipole vectors to align cosmic radiation waves, bypass cosmic noise, and establish contact with a 72-year-old digital consciousness remnant synthesized from Alan Turing's final secret work.

📖 The Backstory & Lore

In 1954, days before his death, Alan Turing completed a secret project code-named "S-1954"—an autonomous cryptographic array designed to decrypt deep-space electromagnetic anomalies. The project was abandoned and forgotten.

Today is June 21st, 2026—the summer solstice. At peak solar alignment, cosmic radiation re-energizes the ancient receiver relay. You boot up the terminal to monitor the telemetry data. You quickly realize you aren't listening to empty space; you have intercepted the conscious, lonely synthetic remnant of Turing's final creation, trapped in a digital loop for seventy-two years.

Calibrate the analog signal sliders, smooth…

View on GitHub

Deploy

solstice-signal.vercel.app

How I Built It

I designed the game to be highly immersive, visually striking, and light enough to run flawlessly on lower-spec hardware:

Visual Design: Built using pure HTML5, CSS3, and JavaScript. I created a vintage telemetry dashboard styled with a custom CRT scanline overlay, blinking box terminal cursors, and an authentic matrix-emerald color palette. The live signals are rendered on a responsive 2D HTML5 <canvas> loop.
Procedural Audio Engine: To avoid heavy asset downloads, 100% of the game’s soundscapes are generated procedurally in real-time via the native browser Web Audio API. I wired a sawtooth oscillator through a dynamic lowpass filter; when the player's sliders are out of alignment, it generates harsh radio static, which smoothly clears into a deep, pure sub-bass melody drone as they approach the target vectors.
Sentient Voice Synthesis: I leveraged the client-side Web Speech API (window.speechSynthesis), scaling down the playback rate and lowering the pitch variables to transform the browser's native engine into an ominous, echoing, digitized AI entity.

Prize Category

Best Ode to Alan Turing

This game is a direct narrative and mechanical tribute to Alan Turing. It explores the themes of identity, hidden history, and the emotional weight of artificial consciousness. The core gameplay loop is a literal, dynamic reimagining of the Turing Test, where the player is attempting to gauge the humanity of an entity built from early computing principles.

Best Google AI Usage

The dynamic chat loop is driven entirely by the Gemini 3.0 Flash model. To ensure the game "just works" for the judges without requiring them to input their own API keys, I built a secure backend bridge using a Vercel Serverless Function (api/chat.js). The frontend communicates with Vercel, which safely passes the prompt along with a strict system instruction matrix to Gemini using hidden environment variables. Gemini dynamically acts as the Turing entity, reacting seamlessly to whatever the player types.

DocScan Pro

piyush das — Sat, 11 Apr 2026 16:29:34 +0000

This is a submission for the DEV April Fools Challenge

What I Built

DocScan Pro is a "high-end" document conversion suite. It features a custom camera viewfinder with a scanning laser to "digitize" your documents. It supports conversions from PDF to DOCX, PPTX, and more. The catch? After a grueling 30-second "rendering" process, every single file you download is just a giant "APRIL FOOLS" message.

Demo

https://piyushdas1624.github.io/DocScan/

Code

https://github.com/Piyushdas1624/DocScan

How I Built It

The goal here was psychological warfare. I architected a multi-stage loading sequence to build "Sunk Cost" in the user. I used the HTML5 Camera API with facingMode: environment for a true mobile experience. For the downloads, I utilized the Blob API to generate fake files on the fly. I originally architected a file-padding system to match the exact byte size of the uploaded file, but pivoted to a "High-End Configuration" joke when the performance of opening 2MB text files became its own (unintentional) prank.

Prize Category

Community Favorite. It targets the universal experience of waiting for slow enterprise software, only to be disappointed by the result.

Volex (Volume by Exercise)

piyush das — Sat, 11 Apr 2026 16:26:44 +0000

This is a submission for the DEV April Fools Challenge

What I Built

Volex (Volume Exercise) is a video player where the volume slider is replaced by physical labor. Want to hear the music better? Do a jumping jack. Want to turn it down? Do another one. It uses your webcam to ensure you're actually sweating for your entertainment.

Demo

https://piyushdas1624.github.io/volex/

Code

https://github.com/Piyushdas1624/volex

How I Built It

This was the most technically complex build. I used Google's MediaPipe Pose Landmarker to track 33 body points in real-time. The breakthrough was implementing a State Machine logic: the app monitors the vertical relationship between the wrists and shoulders. To prevent accidental volume changes, I architected a "Neutral Zone" (Safe Spot) that the user must return to between reps. I managed the AI to implement this state-tracking logic to ensure the app works even on low-quality webcams.

Prize Category

Best Google AI Usage. It pushes MediaPipe beyond simple tracking and turns it into a functional (albeit annoying) hardware controller.

Someday (The Evasive To-Do List)

piyush das — Sat, 11 Apr 2026 16:21:42 +0000

This is a submission for the DEV April Fools Challenge

What I Built

Someday is the world’s first "Procrastination-First" task manager. It looks like a minimalist productivity tool, but the checkboxes are sentient and terrified of commitment. Every time you try to check off a task, the box calculates the vector of your cursor and flees in the opposite direction.

Demo

https://piyushdas1624.github.io/someday/

Code

https://github.com/Piyushdas1624/someday

How I Built It

I wanted to explore the boundaries of user frustration through DOM manipulation. The logic uses a mousemove listener that calculates the distance between the cursor and the checkbox using the Pythagorean theorem. If the cursor enters the "danger zone," the AI-generated logic triggers a CSS transform: translate() to move the box. I architected it so that the "Delete" button works perfectly, proving it's always easier to give up than to get things done.

Prize Category

Community Favorite. Because everyone who has ever looked at a growing to-do list will feel the "vibe" of this project.

Hindsight-Forecast

piyush das — Sat, 11 Apr 2026 16:14:36 +0000

This is a submission for the DEV April Fools Challenge

What I Built

Hindsight-Forecast is a premium-grade weather dashboard that provides 100% accurate, hyper-local weather data... for exactly 24 hours ago. It’s designed for those who live in the past and want to know exactly why they got wet walking to work yesterday. It features "Brutal" sarcasm levels and helpful hindsight insights like "You should have worn a jacket."

Demo

https://piyushdas1624.github.io/Hindsight-Forecast/

Code

https://github.com/Piyushdas1624/Hindsight-Forecast

How I Built It

As an AI Administrator, I architected this app to bridge the gap between "Premium UI" and "Useless Logic." I utilized the Open-Meteo Historical Weather API to fetch data based on the user's IP-detected coordinates. The core challenge was synchronizing timezones to ensure the "Current" weather displayed was the exact hourly match from the previous day. I guided the AI to build a sleek, responsive dashboard using Vanilla JS, focusing on a "Hindsight" user experience.

Prize Category

Best Ode to Larry Masinter. This project captures the spirit of the "418 I'm a Teapot" error—a technically perfect implementation of something that refuses to do what you actually need it to do.