DEV Community

Kushagra Bhardwaj
Kushagra Bhardwaj

Posted on

RE-ALM — Turn Your Real World Into an RPG Adventure 🌍⚔️

Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass Submission 🌿

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

What if going outside felt less like another item on your to-do list and more like starting a new anime adventure?
That's the idea behind RE — an AI-powered, real-world RPG designed to turn everyday activities into quests, progress, and rewards.
Instead of spending every gaming session in a virtual world, RE encourages players to explore the one they already live in. A walk, an outdoor activity, or a small personal challenge can become part of a larger adventure.

The core gameplay loop combines three things:

  • AI-generated quests: An AI Game Master generates structured adventures rather than simply chatting with the player.
  • EXP and progression: Completing quests contributes to player progression, including experience points, levels, and stats.
  • Aether and summons: Players earn Aether through the game's reward system and spend it on summons, with rarity tiers ranging from R and SR to SSR and UR.
  • Inventory and collection: Summoned rewards become part of the player's collection, giving progression a tangible in-game goal.

The idea is to make the phone a launchpad, not the destination. Players should be able to discover their next objective, head outside, complete it, and return to see their progress.

RE is being developed as a phased project, with the current implementation focused on the core gameplay loop: player setup, AI quest generation, quest tracking, progression, the Aether economy, summons, and inventory.

Demo

Live demo / video walkthrough: [https://drive.google.com/file/d/1IZKfCaZ2hCK589dGUSmEYdJ7sf-gPa3j/view?usp=sharing]
The demo should show the complete loop:

  • Open and view the player HUD.
  • Generate an adventure with the AI Game Master.
  • Start and complete a quest using the currently implemented verification flow.
  • Claim the resulting progression rewards.
  • Use Aether to perform a summon and view the result in the collection. I want the demo to show not just the interface, but how the system connects AI-generated content to actual gameplay progression.

Code

GitHub repository: https://github.com/Kush05Bhardwaj/Re-Alm
The project uses a monorepo structure with a Next.js frontend and a Python FastAPI backend. The backend handles gameplay APIs, AI quest generation, progression, the Aether ledger, summons, and inventory, with MongoDB used for persistence.

How I Built It

The main technology choices are:

  • Next.js, TypeScript, and Tailwind CSS for the web interface.
  • Python and FastAPI for the backend APIs and gameplay logic.
  • Gemma through Ollama for open-weight AI inference and AI-generated quests.
  • MongoDB for player data, quests, progression, economy transactions, and collection state.

The important architectural decision was to treat the AI as a game-content generator, not just a chatbot.

The AI Game Master is intended to produce structured quest data that the backend can validate and use in the gameplay system. That data connects to quest tracking, EXP and level progression, Aether rewards, and the summon and inventory systems.

This separation matters: the model proposes the adventure, while application logic manages the player's state and rewards.

The project has been implemented in phases so that the basic gameplay loop can be established before expanding into more advanced systems such as adaptive adventures, NPCs, classes, bosses, and persistent world events.

Why Does Open Innovation Matter?

For RE, open innovation matters because the AI Game Master is one of the core systems of the game.
Using an open-weight model such as Gemma gives the project more flexibility than building the entire experience around a proprietary, hosted model API.

  • First, local inference is an option. With Gemma running through Ollama, the model can be run locally on a suitably configured machine. That creates a path toward experimenting with AI-generated adventures without sending every prompt to a third-party inference service.
  • Second, the model is not a permanent dependency. An open model-based architecture makes it easier to experiment with different models, compare their quest-generation quality, and adjust prompts and output validation as the game evolves.
  • Third, experimentation becomes more accessible. An AI Game Master needs to generate useful, structured content consistently. Being able to experiment locally can make development more practical and avoid relying entirely on per-request hosted inference costs.

Open innovation also comes with real engineering responsibilities. A generated quest must be validated before it affects gameplay, and local inference does not automatically make every part of the application offline or private. RE still needs careful handling of quest safety, reward integrity, and player data.

For me, the goal is to make AI-powered gameplay more customizable and approachable while keeping the actual game rules under the application's control.

What's Next for RE:ALM?

The current implementation covers the foundational gameplay loop through Phase 8, including AI-generated quests, progression, the Aether economy, summons, and inventory. But this is only the beginning.

Here's what's planned next:

  • Adaptive Adventures & Discovery** — quests that adapt to player progress, difficulty, and discoveries.
  • Deeper Immersion** — voice-based, screen-off gameplay, NPC interactions, abilities, and character classes.
  • World Building** — multiple realms, bosses, exploration, fog of war, and dynamic world events.
  • Social & Persistent Worlds** — parties, guilds, leaderboards, an improved economy, and AI-driven storylines.
  • Advanced Immersion** — richer audio experiences, wearables, and future AR/VR exploration.
  • Reliability & Release** — security, privacy, testing, UI polish, mobile/PWA improvements, analytics, and deployment.
  • Documentation & Beyond** — complete documentation, challenge submission, a clearly defined MVP, and the long-term roadmap.

RE:ALM is being built incrementally. The goal isn't to add every feature at once, but to create a reliable core experience before expanding the world.

Final Thought

The world already has enough screens competing for our attention. What if technology helped us engage with the world instead?

RE is my attempt to explore that idea through AI, RPG progression, and real-world quests.

Your reality. Your realm. Your adventure.

Top comments (0)