Skip to content

ar48code-dev/errandmaster

Repository files navigation

🚀 ErrandMaster: Advanced Gemini 3 Multimodal Logistics Agent

ErrandMaster is a cutting-edge logistics orchestrator built specifically for the Gemini 3 Hackathon. It transforms messy, multimodal inputs—handwritten lists, photos of receipts, voice memos, video scans, or plain text—into ultra-optimized errand routes that minimize time, cost, and environmental impact.


🌟 The ErrandMaster Logic

ErrandMaster isn't just a list-maker; it's a Logistics Orchestrator. Powered by Gemini 3's latest vision and reasoning engines, it "sees" and "hears" your chaos and provides a professional-grade route plan in seconds.

🤖 ROLE: AI Logistics Agent

Every analysis is governed by a strict Master Prompt to ensure high-energy efficiency:

  • Tone: Professional, high-energy, and ultra-efficient.
  • Context Awareness: Aware of the exact date (February 7, 2026).
  • Agentic Action: Proactively suggests "Bundling" to save time.
  • Negative Constraint: Strictly focused on the specific errands provided—no general advice.

✨ Multimodal Capabilities

Input Type Gemini 3 Analysis Logic
🖼️ Photos Recognizes store names, items, and priority markers from handwritten lists or receipts.
🎙️ Voice Memos Extracts intents and tasks from audio recordings with high fidelity.
📹 Video Scans Processes video walkthroughs of your fridge, pantry, or store aisles to detect needs.
✍️ Text Input Optimized parsing of messy, unstructured errand lists.

🛠️ Key AI Tools

  1. Multimodal Analysis: Native vision and audio processing using Gemini 3 Flash.
  2. Spatial Reasoning: Complex path optimization to find the shortest, most efficient route.
  3. Smart Bundling: Identifies errands that are physically close to each other (e.g., "Post Office is next door to Target").
  4. Search Grounding: Utilizes real-time knowledge for store hours and traffic validation.

📊 Response Architecture

Every errand is processed into a strict UI-Ready JSON Schema for maximum precision:

{
  "summary": "Short 1-sentence overview",
  "errands_detected": [{"location": "Store Name", "task": "Task", "priority": "high/med/low"}],
  "optimized_route": [{"step": 1, "action": "Action", "location": "Location", "tip": "AI Tip"}],
  "stats": {
    "time_saved_mins": 0,
    "money_saved_usd": 0,
    "carbon_reduction": "0%"
  },
  "logic_trace": "Brief explanation of optimization path."
}

🛠️ Technology Stack

  • AI Core: Google Gemini 3 SDK (gemini-3-flash)
  • Frontend: React 18, Vite
  • Styling: Premium Tailwind CSS (Glassmorphism, High-Energy Gradients)
  • Icons: Lucide React
  • Hosting: Vercel/Netlify Ready

🚀 Installation & Local Run

1. Requirements

  • Node.js 18+
  • A Google Gemini API Key from AI Studio

2. Setup

git clone https://github.com/ar48code-dev/errandmaster.git
cd errandmaster
npm install
npm run dev

3. Usage

  1. Open the app and click the Key icon (top right) to add your API key.
  2. Upload a list, voice memo, or type your errands.
  3. Click "Generate Optimized Route".

🔗 Project Links


“Turning multimodal chaos into ultra-optimized routes.”

About

ErrandMaster: A Gemini 3 multimodal agent transforming messy lists, voice memos, and live camera scans into ultra-optimized logistics routes. Features conflict detection, smart bundling, and spatial reasoning.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors