README.md

---
title: LifeAdmin AI - Autonomous Life Management Agent
emoji: 🤖
colorFrom: purple
colorTo: blue
sdk: gradio
sdk_version: 6.0.1
app_file: app.py
pinned: true
license: mit
tags:
  - mcp-in-action-track-enterprise
  - mcp-in-action-track-consumer
  - mcp-in-action-track-creative
---

# 🤖 LifeAdmin AI - Your Autonomous Life Management Agent

> **🏆 Hugging Face MCP Birthday Hackathon Submission**  
> **Track 2:** MCP in Action 
> **Tags:**
  - mcp-in-action-track-enterprise
  - mcp-in-action-track-consumer
  - mcp-in-action-track-creative

## 📺 Demo Video

**▶️ Watch Demo Video** https://drive.google.com/file/d/1Hdo_hHE-tohYYDmbmX5NJuZY1hrBKcW-/view?usp=sharing


## 📱 Social Media

**LinkedIn Post:** https://www.linkedin.com/posts/maheen-touqeer-3b5b03289_mcp1stbirthday-aiagents-mcp-activity-7400925160662405121-4n97?utm_source=share&utm_medium=member_desktop&rcm=ACoAAEYb6nkB0latOGw2UbEtSUJfOzMNGaP4OPs


## 🎯 Overview

**LifeAdmin AI** is an autonomous AI agent that revolutionizes how professionals manage life administration tasks. Built with Gradio 6 and powered by the Model Context Protocol (MCP), it demonstrates advanced autonomous reasoning, planning, and execution capabilities.

### The Problem

Professionals waste **4-6 hours weekly** on repetitive admin tasks:
- Manual form filling
- Document organization
- Calendar management
- Email composition
- Deadline tracking
- File searching

### Our Solution

An **autonomous AI agent** that:
✅ Understands complex requests through natural language  
✅ Plans multi-step workflows autonomously  
✅ Executes tasks using 7+ MCP tools  
✅ Learns from past interactions via persistent memory  
✅ Provides real-time feedback on reasoning process  

## ✨ Key Features

### 🤖 **Autonomous Agent Capabilities**

#### 1. **Intelligent Planning**
- Analyzes user requests and breaks them into actionable tasks
- Selects appropriate MCP tools for each subtask
- Handles dependencies and sequential execution

#### 2. **Contextual Reasoning**
- Uses RAG (Retrieval-Augmented Generation) to search uploaded documents
- Leverages conversation history from persistent memory
- Makes intelligent decisions based on available context

#### 3. **Multi-Step Execution**
- Executes complex workflows autonomously
- Provides real-time thought traces showing reasoning
- Handles errors gracefully with fallback strategies

### 🔧 **MCP Tool Suite (7 Tools)**

1. **📄 OCR Text Extraction** - Extract text from images and scanned documents
2. **📑 PDF Analysis** - Summarize PDFs and extract metadata (dates, amounts, entities)
3. **📋 Smart Form Filler** - Auto-fill DOCX/XLSX forms with intelligent field matching
4. **✉️ Email Drafter** - Generate context-aware professional emails
5. **📅 Calendar Generator** - Create ICS calendar events from natural language
6. **🗂️ File Organizer** - Organize files by type, date, or size
7. **🔍 RAG Search** - Semantic document search with vector embeddings

### 💬 **Three Interaction Modes**

#### 1. **Manual Dashboard** - Granular Control
Direct access to each MCP tool with visual interface for precise control.

#### 2. **LifeAdmin Coach** - Conversational AI
Chat with an AI assistant that:
- Answers questions about uploaded documents
- Automatically calls appropriate tools
- Provides summarization and insights
- Tracks conversation history

#### 3. **Voice Agent** - Hands-Free Operation
Speak commands naturally and watch the agent:
- Transcribe speech (Groq Whisper)
- Plan and execute tasks
- Respond with synthesized speech (ElevenLabs)

## 🏗️ Architecture

### **Tech Stack**

```
Frontend:  Gradio 6.0 
Agent:     Custom autonomous agent with planning & reflection
MCP:       7 custom MCP tool servers (local fallback mode)
RAG:       ChromaDB + Sentence Transformers (all-MiniLM-L6-v2)
Memory:    SQLite persistent storage
LLMs:      OpenAI GPT-4, Hyperbolic, Groq Llama 3.3  (fallback chain)
Voice:     Groq Whisper (STT), ElevenLabs (TTS)
```

### **Agent Architecture**

```
User Request
    ↓
┌─────────────────┐
│  Agent Core     │
│  - Planning     │ ←→ [Memory Store]
│  - Reasoning    │ ←→ [RAG Engine]
│  - Execution    │
└────────┬────────┘
         ↓
┌────────────────────┐
│   MCP Client       │
│   Tool Router      │
└────────┬───────────┘
         ↓
┌─────────────────────────────────┐
│  MCP Tool Servers (7)           │
│  • OCR  • PDF  • Forms          │
│  • Email  • Calendar  • Files   │
│  • RAG                          │
└─────────────────────────────────┘
```

### **Key Technical Decisions**

1. **Local MCP Execution** - All tools run locally with async fallback for reliability
2. **Multi-Provider LLM** - Graceful degradation across OpenAI → Hyperbolic → Groq
3. **RAG Integration** - Document context enhances agent decision-making
4. **Persistent Memory** - SQLite stores conversations and user preferences
5. **Async-First Design** - Non-blocking I/O for responsive UX

## 🎬 Real-World Use Cases

### **Use Case 1: Automated Form Filling**
```
User uploads membership form + provides data
Agent: Plans extraction → Fills form → Returns completed document
Time saved: 15 minutes → 30 seconds
```

### **Use Case 2: Document Analysis**
```
User: "What are the key deadlines in my PDF?"
Agent: Searches RAG → Extracts dates → Summarizes deadlines
Output: Structured deadline list with context
```

### **Use Case 3: Email Workflow**
```
User: "Draft a follow-up email for my membership application"
Agent: Retrieves application context → Generates professional email → Offers editing
Result: Publication-ready email in seconds
```

### **Use Case 4: Calendar Management**
```
User: "Create a meeting next Tuesday at 2pm"
Agent: Parses natural language → Generates ICS file → Ready for import
Integration: Works with Google Calendar, Outlook, Apple Calendar
```

## 📖 How to Use

### **1. Manual Dashboard**

1. Upload files (PDF, images, documents)
2. Select a tool (OCR, PDF, Forms, etc.)
3. Configure parameters
4. Execute and download results

### **2. LifeAdmin Coach**

1. Upload documents in Manual Dashboard first
2. Switch to Coach tab
3. Ask questions like:
   - "What documents have I uploaded?"
   - "Summarize my PDF"
   - "What's on my calendar?"
   - "Draft an email about X"

### **3. Voice Agent**

1. Upload files (optional)
2. Click microphone or type command
3. Click "Execute Command"
4. Watch agent reasoning in real-time
5. Receive voice + text response

## 🏆 Why LifeAdmin AI?

### **Judging Criteria Alignment**

#### ✅ **Design/Polished UI-UX** 
- Clean, modern interface with Gradio 6
- Responsive mobile design
- Intuitive tab-based navigation
- Real-time feedback and progress indicators
- Professional gradient theming

#### ✅ **Functionality** 
- **Autonomous agents:** Full planning-reasoning-execution cycle
- **MCP integration:** 7 custom MCP tools
- **Advanced features:** RAG, persistent memory, multi-modal I/O
- **Gradio 6 features:** SSR, mobile support, async operations

#### ✅ **Creativity** 
- Novel combination of voice + chat + manual control
- Intelligent form-filling with fuzzy matching
- Thought trace visualization for transparency
- Context-aware tool selection

#### ✅ **Documentation** 
- Comprehensive README with architecture diagrams
- Demo video showing all features
- Clear setup instructions
- Code comments and docstrings

#### ✅ **Real-World Impact** 
- **Target users:** 100M+ professionals worldwide
- **Time savings:** 4-6 hours/week per user
- **Pain points addressed:** Forms, emails, documents, scheduling
- **Scalability:** Cloud-ready architecture

###🚀 **Quick Start**
Try it Online
👉 Open LifeAdmin AI on Hugging Face
https://huggingface.co/spaces/MCP-1st-Birthday/LifeAdmin-AI

Run Locally
bash# Clone repository
git clone https://github.com/MCP-1st-Birthday/LifeAdmin-AI.git
cd LifeAdmin-AI

# Install dependencies
pip install -r requirements.txt

# Set up API keys (at least one required)
export OPENAI_API_KEY="your_key"        
export GROQ_API_KEY="your_key"          
export ELEVENLABS_API_KEY="your_key"    
export HYPERBOLIC_API_KEY="your_key" 

# Launch app
python app.py
Visit http://localhost:7860

## 🤝 Team

**Solo Developer:** Maheen Touqeer  
**HuggingFace:** https://huggingface.co/Maheentouqeer1
**LinkedIn:**  https://www.linkedin.com/in/maheen-touqeer-3b5b03289/


## 🙏 Acknowledgments

Built for **MCP's 1st Birthday Hackathon** hosted by:
- **Anthropic** - For the amazing Claude API and MCP protocol
- **Gradio Team** - For Gradio 6 and excellent documentation
- **Hugging Face** - For hosting infrastructure and community
- **Sponsor APIs** - OpenAI, Groq, ElevenLabs, Hyperbolic

Special thanks to the MCP community for inspiration and support! 🎉


## 📞 Contact & Feedback

**Questions?** Reach out on Hugging Face Space Community

**Found it useful?** ⭐ Star this Space and share with your network!

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference