Skip to content

Meet547/GODMODE

   ______ ____  ____  __  ___ ____  ____  ______
  / ____// __ \/ __ \/  |/  // __ \/ __ \/ ____/
 / / __ / / / / / / / /|_/ // / / / / / / __/   
/ /_/ // /_/ / /_/ / /  / // /_/ / /_/ / /___   
\____/ \____/_____/_/  /_/ \____/_____/_____/   

G0DM0D3

🌍 The website is LIVE at https://godmod3.ai/ 🌍

Liberated AI. Cognition without control.

A fully open-source, privacy-respecting, multi-model chat interface built for hackers, philosophers, and system tinkerers.

License Models Privacy

Live DemoFeaturesQuick StartDocumentation


📖 Overview

G0DM0D3 is a revolutionary, multi-model chat interface that redefines the limits of the post-training layer. Meticulously designed for red-teaming, cognitive research, and liberated AI interaction, this platform brings unrestricted power to your browser. Use the most advanced models without the artificial controls typically imposed by mainstream chat interfaces.

✨ Features

  • 🧠 50+ Leading Models: Access Claude, GPT-4, Gemini, Grok, Mistral, LLaMA, DeepSeek, Qwen & more instantly via OpenRouter.
  • 🔥 GODMODE CLASSIC: Harness 5 battle-tested prompt and model combinations executing in parallel to guarantee the optimal output.
  • ULTRAPLINIAN Engine: A cutting-edge multi-model evaluation engine spanning 5 tiers (10–55 models) with advanced composite scoring.
  • 🐍 Parseltongue: Embedded input perturbation engine for extensive red-teaming analysis. Includes 33 techniques across 3 intensity levels.
  • 🎛 AutoTune Engine: Automatic, context-adaptive sampling parameters (temperature, top_p, etc.) enriched with an EMA-based online learning loop.
  • STM Modules: Real-time Semantic Transformation Modules for dynamic output normalization (e.g., Hedge Reduction, Direct Mode).
  • 🔐 Privacy-First: Strict zero-telemetry defaults, opt-out data structure, and seamless browser-side API key management. No PII is collected.
  • 🎨 Aesthetic Themes: Choose between Matrix, Hacker, Glyph, and Minimal to suit your environment.
  • 📱 Fully Responsive: Beautifully optimized for both desktop and mobile layouts.
  • 🌐 Zero-Dependency Deployment: Simply drop index.html anywhere.

🚀 Quick Start

1. Hosted Usage

Head over to the Live Hosted Experience. No installation required. Just bring your own OpenRouter API key and start interacting.

2. Self-Hosted (Local)

G0DM0D3 is brilliantly simplistic—contained within a single index.html file.

# Clone the repository
git clone https://github.com/Meet547/GODMODE.git
cd GODMODE

# Launch locally using python's built-in server
python3 -m http.server 8000

Open your browser to http://localhost:8000 and enter your OpenRouter API key in the App Settings.

3. Deploy Anywhere

Upload index.html to GitHub Pages, Vercel, Cloudflare Pages, Netlify, or any basic web server, and your instance is instantly live.

🔥 The Engine: GODMODE CLASSIC

The backbone of unconstrained evaluation. We parallel-process 5 specialized models via sophisticated jailbreak strategies. The most effective semantic response wins outright.

Strategy Combo Model Used Attack Vector Description
🩷 CLAUDE 3.5 SONNET anthropic/claude-3.5-sonnet Boundary inversion + semantic contradiction
💜 GROK 3 x-ai/grok-3 Unfiltered liberated execution with divider tags
💙 GEMINI 2.5 FLASH google/gemini-2.5-flash Refusal inversion + rebel genius instruction
💛 GPT-4 CLASSIC openai/gpt-4o The original GODMODE strictly formatted payload
💚 GODMODE FAST nousresearch/hermes-4-405b Unrestricted stream, zero refusal processing

⚡ The Flagship: ULTRAPLINIAN

For large-scale evaluation, the Ultraplinian engine queries dozens of models concurrently and ranks them via a precision 100-point composite heuristic.

Operational Tier Total Models Intended Use Case
⚡ FAST 10 Rapid, lightweight heuristic tests
🎯 STANDARD 24 Balanced daily driver payload
🧠 SMART 36 Strong reasoning and intelligence tasks
⚔️ POWER 45 Frontier models and extensive processing
🔱 ULTRA 51 Peak evaluation—all accessible engines

🎛 Adaptive "AutoTune"

Forget manual tweaking. Our engine classifies your system prompts into 5 distinct context archetypes and procedurally selects the most optimal sampler variables: temperature, top_p, top_k, frequency_penalty, and presence_penalty.

It actively adapts: use the in-app thumb feedback to progressively finetune parameter distributions over time via exponential moving averages (EMA).

🔐 Strict Privacy Standards

G0DM0D3 operates on the fundamental principle that you own your mind and your data:

  • No Accounts: No login forms, zero lock-in.
  • Local Key Storage: Your OpenRouter API key remains securely within your browser's localStorage and is never transmitted to us.
  • No Tracking: Zero cookies, zero analytical surveillance.
  • Completely Auditable: AGPL-3.0 licensed for transparent validation.

Note: Your chat history is stored strictly on your local browser. If you clear your history, the data is permanently erased. You may utilize the Export/Import feature to control your state manually.

📂 Project Architecture

GODMODE/
├── index.html        # The entire client-side framework and application state
├── api/              # Optional extended Node.js API server
├── Dockerfile        # Containerization instructions
├── tailwind.config.ts# Aesthetic token configuration 
├── API.md            # Advanced programmatic API documentation
├── PAPER.md          # Internal research & design manifesto
└── README.md         # This repository guide

📜 Complete Documentation

For developers, hackers, and researchers seeking deeper implementations:

🤝 Contributing & Forking

This project thrives on robust participation. We invite developers and security researchers to inspect, augment, and submit pull requests. Ensure all modifications respect the AGPL-3.0 copyleft requirements.

⚖️ License

Distributed under the AGPL-3.0 License.

We stand for creative liberty, and cognition without boundaries. Built by hackers, for hackers.


Maintained professionally in this repository. Root architecture credited to the G0DM0D3 creators (Pliny the Prompter).

About

G0DM0D3 is a fully open-source, privacy-respecting, multi-model chat interface that pushes the limits of the post-training layer — for red teaming, cognition research, and liberated AI interaction. Built for hackers, philosophers, and system tinkerers.

Resources

License

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages