Get Started in 5 Minutes

The fastest way to start using TAXIA for Korean tax law queries

This guide will get you from installation to your first successful query in just 5 minutes.

⚡ Quick Start Path

Step 1: Install TAXIA (30 seconds)

pip install taxia-core

Expected output:

Successfully installed taxia-core-1.0.1

Step 2: Set Your API Key (30 seconds)

Choose one LLM provider:

Option A: Anthropic Claude (Recommended)

export ANTHROPIC_API_KEY="sk-ant-..."

Option B: OpenAI GPT

export OPENAI_API_KEY="sk-..."

💡 Don't have an API key yet? You can still test with Demo Mode (see Step 3).

Step 3: Run Your First Query (1 minute)

Option A: Demo Mode (No API key required)

Perfect for testing and learning:

from taxia import TaxiaEngine

# Initialize in demo mode
engine = TaxiaEngine(demo_mode=True)

# Ask a question in Korean or English
result = engine.answer("법인세율은 몇 퍼센트인가요?")

# Print the answer
print(f"Answer: {result.answer}")
print(f"Citations: {len(result.citations)} legal sources")

Expected output:

Answer: 법인세율은 과세표준 구간에 따라 다음과 같습니다:
- 2억원 이하: 10%
- 2억원 초과 200억원 이하: 20%
- 200억원 초과 3,000억원 이하: 22%
- 3,000억원 초과: 25%

Citations: 2 legal sources
  1. Corporate Tax Act Article 55 (Tax Rates)
  2. Corporate Tax Act Enforcement Decree Article 92

Option B: Full Mode (API key required)

For real data and production use:

from taxia import TaxiaEngine

# Initialize with your API key (from environment)
engine = TaxiaEngine()

# Query will use real tax law data
result = engine.answer("What is the corporate tax filing deadline?")

print(f"Answer: {result.answer}")
print(f"Trace ID: {result.trace_id}")  # For audit trail

Step 4: Verify Installation (30 seconds)

Check system status:

taxia health

Expected output:

========================================
TAXIA System Health Check
========================================

✅ LLM Provider: Connected (Claude)
⚠  Vector Store: Not configured (Demo mode available)
⚠  Graph Store: Not configured (Optional)

Status: Operational (Demo mode)
========================================

✅ Success! What's Next?

You've successfully:

⚡ JavaScript Quick Start (1 minute)

npm install @xaikorea0/js-client

import { TaxiaClient } from '@xaikorea0/js-client';

const client = new TaxiaClient({
    apiKey: 'YOUR_API_KEY',
    cacheEnabled: true
});

const res = await client.query('법인세율은 얼마인가요?');
console.log(res.answer);

• JS 가이드: api/javascript.html
• React/Vue/Next 예제: src/js/examples/

Choose Your Path:

🎯 Path 1: Keep Testing (Recommended for beginners)

Continue with demo mode to learn TAXIA features: - Try more questions - Explore different query types - Test CLI commands

Next step: Examples

🚀 Path 2: Set Up Production Environment

Ready for real data? Set up the full stack: - Install Qdrant (vector search) - Download tax law data - Index documents

Next step: Production Setup

🔧 Path 3: Build Your Application

Integrate TAXIA into your project: - REST API server - Chatbot integration - Custom workflows

Next step: API Reference

🚀 Production Setup (Optional)

Want to use real tax law data? Follow these steps:

1. Install Qdrant (Vector Database)

Using Docker (Easiest):

docker run -d -p 6333:6333 qdrant/qdrant

Without Docker:

pip install qdrant-client
# Qdrant will run in-memory mode

2. Download Tax Law Data

# Automatic download (recommended)
taxia download --all

# Or specify year
taxia download --year 2025

Data will be saved to: ~/.taxia/data/

3. Index Documents

taxia index ~/.taxia/data/

This will take 2-5 minutes depending on your data size.

4. Query Real Data

from taxia import TaxiaEngine

# Now uses real indexed data
engine = TaxiaEngine()
result = engine.answer("법인세법 제27조의 내용은?")
print(result.answer)

🎓 Learning Path

Beginner Track (Day 1)

1. ✅ You are here - 5 Minute Quick Start
2. Basic Queries - Try different questions
3. CLI Tools - Learn command-line interface

Intermediate Track (Day 2-3)

1. Configuration - Customize settings
2. Data Setup - Understand data structure
3. Advanced Examples - Complex queries

Advanced Track (Week 1+)

1. REST API - Build web services
2. Architecture - Understand system design
3. Production Deployment - Go live

🚨 Troubleshooting

❌ "ModuleNotFoundError: No module named 'taxia'"

Solution:

pip install taxia-core
# Or upgrade to latest version
pip install --upgrade taxia-core

❌ "API key not configured"

Solution:

# Set your API key
export ANTHROPIC_API_KEY="your-key-here"

# Or use demo mode
engine = TaxiaEngine(demo_mode=True)

❌ "Connection refused: Qdrant"

Solution:

Option 1 - Use demo mode:

engine = TaxiaEngine(demo_mode=True)  # No Qdrant needed

Option 2 - Start Qdrant:

docker run -d -p 6333:6333 qdrant/qdrant

Option 3 - Use in-memory mode:

from taxia import TaxiaEngine, TaxiaConfig

config = TaxiaConfig(qdrant_host=None)  # In-memory
engine = TaxiaEngine(config=config)

❌ Still having issues?

1. Check Troubleshooting Guide
2. Visit GitHub Issues
3. Review Documentation

📊 Demo Mode vs Production Mode

💡 Quick Tips

Tip 1: Start with Demo Mode

Don't have API keys or Qdrant ready? No problem! Demo mode is perfect for: - Learning TAXIA features - Testing integration - Proof of concepts - Training and demos

Tip 2: Use Korean Questions

TAXIA is optimized for Korean tax law:

result = engine.answer("근로소득 공제액은 얼마인가요?")

Tip 3: Check Audit Trail

Every query has a unique trace ID:

print(f"Trace ID: {result.trace_id}")
# Use this for debugging and compliance tracking

Tip 4: Explore CLI Tools

Quick health checks and configuration:

taxia health         # System status
taxia config show    # View settings
taxia --help         # All commands

🎉 Next Steps

Congratulations! You're now ready to use TAXIA.

Recommended next actions:

1. Try more examples: Examples Page
2. Learn configuration: Config Guide
3. Build something: Use Cases

Quick Example Ideas:

Tax Chatbot:

while True:
    question = input("Ask a tax question: ")
    result = engine.answer(question)
    print(f"\n{result.answer}\n")

Batch Processing:

questions = [
    "법인세 신고기한은?",
    "부가가치세율은?",
    "근로소득세 계산법은?"
]

for q in questions:
    result = engine.answer(q)
    print(f"Q: {q}")
    print(f"A: {result.answer}\n")

API Server:

taxia server --port 8000
# Visit http://localhost:8000/docs

⭐ Success Stories

"5분 만에 세무 챗봇 프로토타입 완성!"
- Developer at FinTech Startup

"Demo 모드로 먼저 테스트 후 프로덕션 도입"
- Tax Consulting Firm

"설치부터 첫 쿼리까지 정말 5분 걸렸어요"
- AI Engineer

📚 Additional Resources

• Documentation: Full Docs
• GitHub: Source Code
• PyPI: Package
• Data: Hugging Face

Ready to build amazing tax AI applications? Let's go! 🚀