DEV Community: Aswin Behera

ASCII Snake: Web Crawler

Aswin Behera — Mon, 15 Sep 2025 06:51:39 +0000

This is a submission for the Google AI Studio Multimodal Challenge

What I Built

I created ASCII Snake: Web Crawler, a thematic reimagining of the classic Snake game for the modern web. This isn't just about growing a line on a grid; it's an interactive commentary on the battle for focus in an age of digital distraction.

In the game, you pilot a "snake" that represents a web crawler navigating deeper into a digital system, for this example: Learning Computer Science from Wikipedia.

Goal: Consume data packets (*), which increase your depth (score) and lengthen your URL path, simulating a journey into more profound knowledge.
Obstacles: Encounter X — pesky ads and short-form video temptations.
Distraction Mechanic: Hitting one triggers a pop-up window with an AI-generated image mimicking clickbait and memes. This collision:
- Diverts your attention.
- Penalizes your depth.
- Forces you to refocus your journey.

The entire experience is wrapped in a CLI and ASCII aesthetic, with:

A dynamic data-stream background.
Procedural retro sound effects.
A deeply immersive, old-school hacker ambiance.

How I Used Google AI Studio

Google AI, specifically the Gemini API, powers the game's core “distraction” mechanic. A commentary on the increasing ai-generated content on the web, plus how gemini's guardrails on the latest gemini-2.5 flash image-preview are so good!!!

I used imagen-4.0-generate-001 to generate images on the fly.
This transformed what could have been repetitive obstacles into a dynamic, unpredictable element.

Implementation

When a player hits an obstacle or loses a life, the game triggers an API call to Gemini.
A curated list of text prompts produces images styled as:
- Viral memes.
- Clickbait thumbnails.
- Bizarre, attention-grabbing media.
- Example: > "Clickbait thumbnail, a person with a shocked face, giant red arrow pointing to something mundane"
The API response — a freshly generated image — is rendered into a pop-up window on the game screen.

This makes the AI-generated content a central part of the gameplay loop, directly challenging the player.

Multimodal Features

The key multimodal feature is real-time, prompt-based image generation that responds to in-game events.

Instead of using static pre-loaded assets, the game leverages Gemini’s text-to-image capabilities to create unique visuals for every encounter.

Why This Matters

Authentic Distractions

The novelty of each distraction feels real. Like scrolling a social media feed, players never know what absurd image will appear next.
Infinite Replayability

Because images are generated live, every session is unique. Players won’t become desensitized to repeating content.
Thematic Reinforcement

The AI-generated images embody the narrative: bizarre, surreal, and algorithmically optimized — mirroring modern digital media designed to hijack attention.

Summary

ASCII Snake: Web Crawler takes a classic arcade formula and retools it into a playable critique of distraction culture, where AI-generated clickbait doesn’t just decorate the experience — it is the challenge.

Isekai Chronicles: A Local Neighbourhood Adventure

Aswin Behera — Mon, 15 Sep 2025 05:31:34 +0000

This is a submission for the Google AI Studio Multimodal Challenge

Demo link: Isekai Chronicles: A Local Neighbourhood Adventure

Video walkthrough: My neighbourhood adventure

What I Built

At its heart, Isekai Chronicles is an attempt to answer a simple question:

What if your walk to the corner store was an epic quest? What if the local park was a haunted forest, and the library a repository of ancient lore?

I built this app to solve the problem of static, predictable game worlds. It creates a deeply personal and endlessly replayable text-based RPG by transmuting the player's real-world surroundings into a high-fantasy adventure.

This isn't just a game; it's a generative world engine. Every adventure is unique, woven from the fabric of your own neighborhood and brought to life by a suite of AI-powered tools.

Core Features:

Real-World to Fantasy World
Using your device's geolocation, the app identifies nearby parks, landmarks, and points of interest, using them as the seed to generate a bespoke fantasy setting.
AI Storyteller & Dungeon Master
The narrative is driven by a dynamic AI that writes the story, describes fantastical scenes, invents deep lore on the fly, and presents meaningful choices that shape your journey.
AI Illustrator
Every location, character, and item is visualized with a unique, AI-generated image in a consistent *hand-drawn fantasy ink* style. The world isn't just described; it's sketched out before your eyes.
Cinematic Combat
Enemies don't spawn on a boring battle screen. Instead, the AI dynamically edits the current scene, adding foes and altering your character's pose to create a seamless, confrontational tableau.
On-the-Fly Artistry
Loot items like a *Runic Key* or a *Shimmering Health Potion* are instantly given custom inventory icons by the AI artisan.
Voice-Powered Immersion
Speak your commands directly to the game. The AI listens, transcribes, and advances the story, making the experience more accessible and immersive.
The Adventurers' Archive
At the end of your quest, save your story to a global archive for other players to discover.
Shared World Map
Opt-in to place a pin on the World Map, showing the approximate (privacy-safe) region of your tale, creating a tapestry of shared adventures.
Physical Memento
Download every completed adventure as a beautifully formatted PDF — a true *chronicle* of your personal journey.

How I Used Google AI Studio

Google's Gemini models are the heart, soul, and imagination of this project. I treated each capability like a specialized member of a creative team.

gemini-2.5-flash — The Dungeon Master

Creative Writing: Initial story, branching choices, combat sequences.
Structured Data Generation: Returns clean JSON with game state (choices, loot, enemy stats, lore).
Audio Transcription: Converts spoken commands into text.
Auxiliary Content: Generates fantasy-style regional names for the World Map.

gemini-2.5-flash-image-preview — The Master Illustrator

Generates 16:9 landscape scenes.
Handles narrative-driven edits, e.g., adding enemies during combat.
Central to the cinematic combat system.

imagen-4.0-generate-001 — The Artisan

Produces high-quality 1:1 assets where detail matters.
Creates the player’s character silhouette and custom inventory icons.

Orchestration: A single player choice may trigger:

Story text from gemini-2.5-flash.
Scene edits via gemini-2.5-flash-image-preview.
Item icon creation via imagen-4.0-generate-001.

Multimodal Features

The magic of Isekai Chronicles comes from seamless blending of text, images, and audio — each in conversation with the others.

Text-to-Image — Visualizing the Narrative

Example:

Storyteller describes a “Sun-drenched plaza with a cracked marble statue.”

This description is passed to the Illustrator, producing a matching visual scene.

Result: Perfect cohesion between what you read and what you see.

Image-and-Text-to-Image — Cinematic Combat Editing

Example:

Current scene + instruction: “Add a menacing Stone Golem, and change the character’s pose to be confrontational.”

The Illustrator modifies the existing image instead of replacing it.

Result: A fluid, cinematic transition from exploration to combat.

Audio-to-Text — Speaking Your Reality

Player speaks into the mic.
Audio is base64-encoded and sent to gemini-2.5-flash.
Model transcribes it into a clean command (e.g., “Examine the glowing runes”).

Result: Interaction feels natural — less like clicking buttons, more like speaking your will into existence.

That's Isekai Adventures for you!

How I Automated Competitor Intel for My Indie Game Using n8n, Bright Data, and Notion

Aswin Behera — Mon, 01 Sep 2025 06:57:27 +0000

This is a submission for the AI Agents Challenge powered by n8n and Bright Data

What I Built

As a solo indie developer working on my game serverBound, I often felt lost in the noise:

How do I know what players are saying about games similar to mine?
Which new releases on Steam are resonating, and why?
How do I keep track of this without burning all my creative energy on spreadsheets and manual browsing?

So I built an automated Competitor Intelligence Agent.

This workflow pulls in popular new releases from Steam, enriches them with player counts and review data, sends them to an AI agent for analysis, and finally saves everything neatly into Notion pages — one per competitor game.

The result: a live “Competitor Radar” that tells me what people are praising, what they’re complaining about, and why certain games are hot right now — all without me having to trawl through reviews manually.

Demo

Here’s a snapshot of what lands in my Notion workspace:
Workspace live at: GameDevKernel

Game Page Title: e.g. PEAK
Metadata: Steam App ID, release date, player count, review score, popularity rank.
AI Analyst Summary:
1. What players are most likely talking about
2. Risks/complaints surfacing in reviews
3. Why the game resonates now

n8n Workflow

I published my workflow JSON as a Gist here:

👉 GitHub Gist: Competitor Radar Workflow

Technical Implementation

Step 1 — Fetch new releases

Using the Bright Data Verified Node, I scraped the Popular New Releases section on Steam. This gave me a JSON of games with title, steamAppId, releaseDate, and steamLink.

Step 2 — Enrich with player + review data

For each game, I fanned out two HTTP requests:

Steam’s GetNumberOfCurrentPlayers endpoint → current player count
Steam’s appreviews endpoint → review totals, positives/negatives, and score descriptors

I then merged these by array position into one clean JSON per game.

Step 3 — Rank games by popularity

I wrote a Function node to compute a custom popularity score (player count × log(review volume)) and sorted games into a ranked list.

Step 4 — AI Agent Analysis

Each game summary was piped into an AI Agent Node (Gemini 1.5/2.5 Flash) with this prompt:

“As an AI analyst, tell me: 1. What are players most likely talking about? 2. Any risks or complaints? 3. Why is this game resonating right now?”

The AI produced Markdown-formatted competitive insights.

Step 5 — Save to Notion

Using Notion’s API:

Each game becomes its own page under a parent “Competitor Radar” page.
The page includes all metadata at the top.
The AI’s markdown is split into multiple paragraph blocks (to avoid Notion’s 2000-char limit).

This means I can browse competitor insights directly in my Notion hub alongside my own dev notes.

Bright Data Verified Node

The Bright Data Verified Node was critical here:

SteamDB and some other sources blocked requests outright.
With Bright Data, I could scrape Steam’s “Popular New Releases” reliably.
I also used it to pull game listings from ThinkyGames.com (to expand my competitor pool beyond AAA).

Without Bright Data’s reliability, this workflow wouldn’t have worked.

Journey

I started this project thinking it would just be a quick “scrape Steam reviews” experiment. It ballooned into a full Notion-powered competitor radar.

Challenges I ran into:

Notion API quirks: Had to split large AI outputs into multiple paragraph blocks due to the 2000-character limit.
Merging API results: Steam’s APIs return data in different shapes, so I had to align everything via Merge nodes.
Old model hallucinations: Gemini sometimes assumed release dates were in the future, so I had to post-process context in my Function node.
Staying indie-lean: I resisted the temptation to over-engineer and focused on one high-ROI workflow.

What I learned:

As a solo dev, automation is a force multiplier. I now get daily competitive insights while focusing my energy on making my game.
Notion as a control tower is underrated. It’s not just notes; it can be a living database of competitor intel, player sentiment, and creative research.
n8n + Bright Data + AI is basically a Swiss army knife for game developers.

Why This Matters

For me, this workflow is like having a personal market analyst who works 24/7:

I know exactly what’s resonating with players.
I see early warnings of complaints or pitfalls.
I can benchmark ServerBound against current trends on Steam.

As a single dev, this saves me dozens of hours each week and keeps me focused on what matters: building a game players will love.

✨ That’s my submission! Indie game dev is already hard — but with the right automations, you don’t have to fight blind.

Notion

A tool that connects everyday work into one space. It gives you and your teams AI tools—search, writing, note-taking—inside an all-in-one, flexible workspace.

healthy-surfboard-e7c.notion.site