Domain

GameForge AI

GameForge turns your game idea into a full design document through four specialized agents: 1. **Story Agent** — narrative, characters, world-building, and plot progression 2. **Gameplay Agent** — core loops, progression systems, controls, and multiplayer design 3. **Visuals Agent** — art style guide, aesthetics, VFX, and audio direction 4. **Tech Agent** — engine recommendations, platform requirements, pipeline, and budget estimates Agents hand off sequentially, sharing summaries before producing detailed design sections.

Python Streamlit

← Back to marketplace

GameForge AI

Multi-agent game design team by Sharad Khare.

GameForge AI is a four-agent AutoGen swarm that collaborates on complete game concepts — story, gameplay mechanics, visual/audio direction, and technical recommendations — from your creative brief in a Streamlit interface.

Built and maintained by Sharad Khare — AI strategist, full-stack developer, and creator of practical agent workflows.

What this project does

GameForge turns your game idea into a full design document through four specialized agents:

1. Story Agent — narrative, characters, world-building, and plot progression

2. Gameplay Agent — core loops, progression systems, controls, and multiplayer design

3. Visuals Agent — art style guide, aesthetics, VFX, and audio direction

4. Tech Agent — engine recommendations, platform requirements, pipeline, and budget estimates

Agents hand off sequentially, sharing summaries before producing detailed design sections.

Why Sharad Khare built GameForge AI

Game design spans narrative, mechanics, art, and engineering — rarely handled well by a single prompt. GameForge demonstrates how an AutoGen swarm with role-specific agents produces richer, more coherent game concepts than a one-shot LLM call.

How it works

Game Brief (vibe, type, audience, budget, mechanics...)
   │
   ▼
Story Agent → summary → Story Design section
   │
   ▼
Gameplay Agent → summary → Gameplay Design section
   │
   ▼
Visuals Agent → summary → Visual Design section
   │
   ▼
Tech Agent → summary → Technical Recommendations section
   │
   ▼
Streamlit expanders with full design document

Agent handoff flow

| Agent | Focus | Output |

|-------|-------|--------|

| Story | Narrative and world-building | ## Story Design |

| Gameplay | Mechanics and systems | ## Gameplay Design |

| Visuals | Art and audio direction | ## Visuals Design |

| Tech | Engine, pipeline, constraints | ## Tech Design |

Features

Four-agent AutoGen swarm with sequential handoffs
Rich Streamlit form: genre, audience, platforms, budget, mechanics, mood
Context sharing between agents via swarm summaries
Expandable output sections for each design pillar
Configurable detail level (Low / Medium / High)
Inspiration and unique feature inputs

Quick start

cd gameforge-ai
pip install -r requirements.txt
streamlit run app.py

Required API key

OpenAI API key — enter in the sidebar when running the app

Example inputs

Vibe: Epic fantasy with dragons
Type: RPG · Goal: Save the kingdom from eternal winter
Audience: Young Adults · Art: Stylized · Platforms: PC, Nintendo Switch
Mechanics: Combat, Exploration, Crafting · Mood: Epic, Mysterious
Budget: $50,000 · Timeline: 18 months

Project structure

gameforge-ai/
├── app.py
├── gameforge/
│   ├── config.py
│   ├── prompts.py
│   ├── task_builder.py
│   └── swarm.py
├── requirements.txt
├── pyproject.toml
└── README.md

Topics

Sharad KhareGameForge AIgame design agent teamAutoGen swarm game designAI game concept generatorstory gameplay visuals tech agentsindie game design AIgame mechanics generatorStreamlit game design toolmulti-agent game development.

Who this is for

Indie game developers prototyping concepts quickly
Game design students learning multi-agent workflows
Creative directors exploring AI-assisted ideation
Developers building AutoGen swarm patterns

Use cases

Generate a full game design doc from a creative brief
Explore story, mechanics, art, and tech directions in parallel
Prototype game pitches with budget and platform constraints
Learn how context-sharing swarms improve multi-domain output

Author

Website: sharadkhare.in

License

Loading file tree…

Select a file to preview its source.

bash

cd gameforge-ai
pip install -r requirements.txt
streamlit run app.py