Skip to content

Repository files navigation

GOD MODE: 75-Model Autonomous AI Swarm System

GOD MODE is an autonomous multi-agent swarm platform that orchestrates 75 AI models simultaneously to decompose complex user prompts into parallel specialist subtasks, execute them concurrently across zero-cost cloud providers, synthesize deliverables, and certify output integrity via automated QA verification.

The platform features a clean Anthropic Claude and OpenAI ChatGPT inspired user interface paired with a real-time prompt traversal engine that visualizes data packet flows and neural routing across the swarm mesh with millisecond telemetry.


Key Features

1. Claude & ChatGPT Inspired Professional Interface

  • Refined Design System: Zinc and charcoal dark mode with warm linen light mode, crisp borders, and subtle terracotta and emerald accents.
  • Segmented Navigation: Seamless switching between Chat, Swarm Network, Model Registry, and API Keys & Settings.
  • Centered Conversational Feed: Clean chat column with starter directive cards, structured prose, syntax-highlighted code blocks with 1-click copy, and a floating composer capsule.
  • Collapsible Reasoning Drawer: Step-by-step transparency into the multi-agent pipeline displaying agent roles, model IDs, cloud providers, and individual execution latencies.

2. Real-Time Swarm Network & Prompt Traversal Engine

  • Interactive 75-Node Mesh: 4-column neural layout with dynamic bezier curves connecting Orchestrators, Coders, General/NLP specialists, and Vision/Media agents.
  • 4-Stage Traversal Monitor:
    • Stage 1: Orchestrator Ingest — Highlights the active Brain Orchestrator node decomposing the user's directive.
    • Stage 2: Neural Synapse Route — Launches glowing SVG packet particles across calculated synapse pathways.
    • Stage 3: Specialist Execution — Illuminates the exact coder, NLP, vision, and reasoning specialist nodes assigned to parallel subtasks.
    • Stage 4: Synthesis & QA — Validates the final verified presentation and logs latency telemetry.
  • Live Synapse Stream: Millisecond-precision telemetry feed logging ingestion, subtask allocation, execution latencies, and verification status.
  • Interactive Transmit Bar: Direct prompt transmission bar inside the network canvas to test neural routing in real time.
  • Bidirectional Sync: Prompts submitted in either the Chat tab or Network tab run through the swarm, animate on the network mesh, and record deliverables in the chat session.

3. 100% Zero-Cost Cloud Provider Integrations

GOD MODE connects to 4 free API providers with zero subscription fees required:

  • Google Gemini API (aistudio.google.com): Flagship reasoning, 1M token context, and orchestration with Gemini 2.5 Flash and Gemma 4 (15 RPM / 1,500 RPD free).
  • Groq Cloud API (console.groq.com): High-speed LPU inference running Qwen 2.5 Coder, Llama 3.3 70B, and Compound MoE (30 RPM free).
  • OpenRouter API (openrouter.ai): Access to 70+ free open-source models including DeepSeek R1, Minimax M3, and Mistral.
  • Hugging Face Token (huggingface.co): Multimodal diffusion image generation (FLUX.1 Schnell, Stable Diffusion XL) and BERT NLP transformers.

4. Smart Cross-Provider Cascading Fallback

  • If any provider encounters quota limits or rate throttling (HTTP 429), the swarm dispatcher automatically failovers across available providers in sub-seconds without interrupting the user request.

Multi-Agent Swarm Pipeline

[ User Prompt ]
       │
       ▼
┌─────────────────────────────────────────┐
│  Stage 1: Chief Brain Orchestrator      │
│  - Analyzes directive & context         │
│  - Decomposes into 1-4 modular subtasks │
└─────────────────────────────────────────┘
       │
       ▼
┌─────────────────────────────────────────┐
│  Stage 2: Parallel Specialist Execution │
│  ├── Agent 1: Coder Specialist          │
│  ├── Agent 2: NLP & Translation         │
│  ├── Agent 3: Creative & Narrative      │
│  └── Agent 4: Multimodal AI Vision      │
└─────────────────────────────────────────┘
       │
       ▼
┌─────────────────────────────────────────┐
│  Stage 3: Presentation Synthesizer      │
│  - Compiles raw specialist outputs      │
│  - Preserves full syntax & code blocks  │
│  - Rehydrates visual artifacts          │
└─────────────────────────────────────────┘
       │
       ▼
┌─────────────────────────────────────────┐
│  Stage 4: QA Verifier & Auditor         │
│  - Audits answer completeness           │
│  - Validates code integrity & formatting│
└─────────────────────────────────────────┘
       │
       ▼
[ Final Verified Deliverable ]

75-Model Registry Matrix

Specialist Column Agent Count Primary Model IDs Core Responsibilities
1. Orchestration 15 Models gemini-2.5-flash, groq/compound, allam-2-7b, gemini-3.6-flash, minimax-m3 Directive decomposition, subtask planning, cross-agent coordination
2. Coding 20 Models qwen-2.5-coder-32b, qwen3.8-27b, north-mini-code, laguna-s-2.1, gpt-oss-20b Production-grade software engineering, API design, debugging, algorithms
3. Reasoning & NLP 25 Models llama-3.3-70b, gemma-4-31b, deepseek-r1, glm-5.2, ling-3.0-flash Multilingual translation, sentiment analysis, mathematics, deep reasoning
4. Vision & Media 15 Models black-forest-labs/FLUX.1-schnell, stabilityai/sdxl, fal-ai/fast-svd Multimodal artwork generation, visual diagramming, video prompt direction

Installation & Setup

Prerequisites

  • Python 3.10 or higher
  • Node.js 18+ (Optional, for Electron desktop app)

1. Clone the Repository

git clone https://github.com/B-Sheikh/GOD_MODE.git
cd GOD_MODE

2. Set Up Virtual Environment & Dependencies

# Create virtual environment
python3 -m venv venv

# Activate virtual environment
source venv/bin/activate  # Linux / macOS
# or: venv\Scripts�ctivate  # Windows

# Install required dependencies
pip install -r requirements.txt

3. Configure API Credentials (Zero-Cost)

Create a .env file from the provided example:

cp .env.example .env

Edit .env and add at least one free API key (or enter keys directly inside the Settings tab in the UI):

GEMINI_API_KEY=your_gemini_key_here
GROQ_API_KEY=your_groq_key_here
OPENROUTER_API_KEY=your_openrouter_key_here
HUGGINGFACE_API_TOKEN=your_hf_token_here

4. Run the Application

Option A: Web Browser Mode (FastAPI)

python main.py

Open http://localhost:8000 in your web browser.

Option B: Electron Desktop App

# In terminal 1: Start backend
python main.py

# In terminal 2: Start Electron desktop client
cd jarvis-desktop
npm install
npm start

API Endpoints Reference

Method Endpoint Description
POST /godmode Main swarm orchestrator. Accepts {"prompt": "..."} and returns verified deliverable with full task telemetry.
GET /api/status Returns system diagnostic status, provider connectivity, and masked key states.
GET /api/models Returns the complete catalog of 75 registered AI models with category and provider tags.
GET /api/keys Returns key configuration status and preview values.
POST /api/keys Updates API credentials, writes to .env, and hot-reloads the active swarm cluster.
POST /api/keys/test Verifies live connectivity for a specific provider (gemini, groq, openrouter, huggingface).

Project Structure

GOD_MODE/
├── api_clients.py          # Unified API dispatcher & cross-provider cascading fallback router
├── main.py                 # FastAPI backend, static asset server, and REST API routes
├── models_registry.py      # Registry mapping 75 models across 4 specialist columns
├── orchestrator.py         # Chief Brain Orchestrator prompt decomposition & parallel task runner
├── synthesizer.py          # Multimodal presentation synthesizer & code assembler
├── verifier.py             # QA verification agent & output compliance auditor
├── test_run.py             # End-to-end integration test script
├── requirements.txt        # Python package dependencies
├── .env.example            # Sample zero-cost environment configuration template
├── .gitignore              # Git ignore rules for virtual environments and credentials
├── LICENSE                 # MIT License
└── jarvis-desktop/         # Frontend user interface (Web & Electron Desktop)
    ├── index.html          # Semantic HTML structure, tabs, live HUD, and modals
    ├── styles.css          # Claude & ChatGPT design tokens, typography, and SVG synapse animations
    ├── renderer.js         # Real-time prompt traversal engine, network graph, and chat handlers
    ├── main.js             # Electron main process & IPC window handlers
    └── package.json        # Electron package definition

License

This project is licensed under the MIT License. See the LICENSE file for details.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages