Cost Katana Examples

Production-ready examples for every Cost Katana feature

Overview

This repository contains 350+ production-ready code examples demonstrating how to use Cost Katana across 49 different features. Whether you're building with TypeScript, Python, or directly via HTTP APIs, you'll find working examples with full error handling and best practices.

What you'll find here:

• 🟦 TypeScript/Node.js - Complete SDK examples with type safety
• 🟨 Python - Pythonic examples with the Cost Katana SDK
• 🟩 HTTP REST APIs - Direct API calls you can test immediately
• 🔧 Framework Integrations - Express, Next.js, Fastify, NestJS, FastAPI

---

Quick Start

1. Get Your API Key

Sign up at costkatana.com/dashboard to get your API key.

2. Choose Your Language

TypeScript/Node.js:

npm install cost-katana
export COST_KATANA_API_KEY=your_key_here
npx ts-node 1-cost-tracking/npm-package/openai.ts

Python:

pip install costkatana
export COST_KATANA_API_KEY=your_key_here
python 1-cost-tracking/python-sdk/openai.py

HTTP (REST Client):

1. Install REST Client for VS Code
2. Open any .http file
3. Update the API key
4. Click "Send Request"

---

Examples by Category

🏗️ Advanced Architecture & Intelligence

Leverage Cost Katana's AI-native architecture with autonomous optimization and self-improving systems.

Feature	Description	Examples
Provider-Independent Core	Capability-based routing, zero vendor lock-in	Guide, Patterns
AI Governance & Security	Zero-trust agent governance with RBAC	HTTP, TypeScript, Python
Data Network Effects	Self-improving AI with learning loops	HTTP, TypeScript, Python

Key Benefits:

• AI-Native Architecture - Platform designed from ground up with AI models as first-class infrastructure
• Autonomous Optimization - Hands-free cost reduction with intelligent interventions
• Self-Improving - System gets smarter with every user interaction
• Enterprise Security - Zero-trust governance with comprehensive auditing

---

💰 Cost Management

Track and optimize your AI spending with comprehensive cost management tools.

Feature	Description	Examples
Cost Tracking	Monitor costs across all AI providers	HTTP, TypeScript, Python
Analytics	Analyze spending patterns and trends	HTTP, TypeScript, Python
Budgets	Set spending limits and alerts	HTTP, TypeScript, Python
Cost Optimization	AI-powered cost reduction recommendations	HTTP, TypeScript, Python

Expected Savings: 30-80% cost reduction through optimization

---

🚀 Performance & Optimization

Boost performance and reduce costs with intelligent caching and optimization.

Feature	Description	Examples
Semantic Caching	Cache similar requests (30-40% savings)	HTTP, TypeScript, Python, Frameworks
Cortex Optimization	AI-powered prompt compression	HTTP, TypeScript, Python
Model Routing	Route to optimal models automatically	HTTP, TypeScript, Python
Failover	Auto-failover between providers	HTTP, TypeScript, Python

Performance Gains: 3-10x faster responses with caching

---

🔔 Monitoring & Alerts

Stay informed with real-time notifications and comprehensive monitoring.

Feature	Description	Examples
Webhooks	Real-time event notifications (65+ events)	HTTP, TypeScript, Python, Frameworks
OpenTelemetry	Distributed tracing & metrics	HTTP, TypeScript, Python, Frameworks
Monitoring	System health & performance	HTTP, TypeScript, Python
Alerts	Custom alert rules	HTTP, TypeScript, Python
Email Tracking	Automated email notifications	HTTP, TypeScript, Python

Coverage: 65+ webhook event types available

---

🤖 AI Orchestration

Build complex AI workflows and multi-agent systems.

Feature	Description	Examples
Agent Trace	Monitor and optimize agent traces	HTTP, TypeScript, Python, FastAPI
Agents	AI agent management	HTTP, TypeScript, Python
Memory	Persistent agent context	HTTP, TypeScript, Python
Templates	Reusable prompt templates	HTTP, TypeScript, Python

Capabilities: Parallel execution, conditional branching, error handling

---

🔒 Security & Compliance

Protect your AI applications with built-in security features.

Feature	Description	Examples
Key Vault	Secure API key management	HTTP, TypeScript, Python
Guardrails	Content moderation & PII detection	HTTP, TypeScript, Python
Security	Threat detection & scanning	HTTP, TypeScript, Python
MFA	Multi-factor authentication	HTTP, TypeScript, Python
Audit Logs	Complete audit trail	HTTP, TypeScript, Python

Protection: PII detection, toxicity filtering, prompt injection defense

---

🔧 Developer Tools

Essential tools for development and testing.

Feature	Description	Examples
Gateway	Unified API for all providers	HTTP, TypeScript, Python
CLI	Command-line interface	Shell scripts, Guides
Python SDK	Complete Python package	Python examples
CKQL	SQL-like query language	HTTP, TypeScript, Python
Notebooks	Interactive analysis	HTTP, TypeScript, Python

---

📊 Analytics & Reporting

Comprehensive analytics and automated reporting.

Feature	Description	Examples
Reports	Automated report generation	HTTP, TypeScript, Python
Feedback	Request quality ratings	HTTP, TypeScript, Python
Tagging	Custom request tagging	HTTP, TypeScript, Python
Experiments	A/B testing & comparison	HTTP, TypeScript, Python
AI Cost Monitoring	Real-time cost tracking	HTTP, TypeScript, Python

---

👥 Team & Organization

Collaborate effectively with team management features.

Feature	Description	Examples
Projects	Project organization	HTTP, TypeScript, Python
Team Management	Team collaboration	HTTP, TypeScript, Python
User Telemetry	Usage analytics	HTTP, TypeScript, Python
Rate Limiting	API rate controls	HTTP, TypeScript, Python

---

🔌 Integrations

Connect with your favorite tools and platforms.

Feature	Description	Examples
Integrations	Third-party connections	HTTP, TypeScript, Python
Ingestion	Data import pipelines	HTTP, TypeScript, Python
Express.js	Express integration	TypeScript
Next.js	Next.js integration	TypeScript
FastAPI	FastAPI integration	Python

---

Complete Feature List

All 44 features with direct links to examples:

📋 Click to expand full feature index

Core Features (1-9)

1. Cost Tracking - Track AI costs across providers
2. Gateway - Unified API gateway
3. Optimization - General optimizations
4. Cortex - Prompt optimization
5. Analytics - Cost analytics
6. Advanced - Advanced features
7. Frameworks - Framework integrations
8. Python SDK - Python examples
9. CLI - Command-line interface

Infrastructure (10-17)

10. Webhooks - Real-time notifications
11. Observability - OpenTelemetry & tracing
12. Guardrails - Content moderation
13. Workflows - Multi-step orchestration
14. Cache - Semantic caching
15. Key Vault - Secure key management
16. Security - Threat detection
17. Monitoring - System health

Management (18-24)

18. Projects - Project management
19. Budgets - Budget tracking
20. Memory - Agent memory
21. Templates - Prompt templates
22. Feedback - Request feedback
23. Tagging - Custom tagging
24. Experiments - A/B testing

Specialized Features (25-43)

25. Notebooks - Interactive notebooks
26. CKQL - Query language
27. Agents - AI agents
28. MFA - Multi-factor auth
29. Moderation - Content moderation
30. Cortex Training - Custom training
31. Email Tracking - Email notifications
32. AI Cost Monitoring - Real-time monitoring
33. Ingestion - Data ingestion
34. User Telemetry - Telemetry config
35. Team Management - Team collaboration
36. Audit Logs - Audit trail
37. Rate Limiting - Rate controls
38. Failover - Provider failover
39. Model Routing - Intelligent routing
40. Cost Optimization - Recommendations
41. Alerts - Alert management
42. Reports - Report generation
43. Integrations - Third-party integrations

---

Framework Integration Examples

Express.js

import express from 'express';
import { CostKatana } from 'cost-katana';

const app = express();
const katana = new CostKatana({ apiKey: process.env.COST_KATANA_API_KEY });

app.post('/chat', async (req, res) => {
  const response = await katana.chat.completions.create({
    model: 'gpt-4',
    messages: req.body.messages
  });
  res.json({ 
    response: response.choices[0].message, 
    cost: response.cost 
  });
});

See: 7-frameworks/express/

Next.js API Route

// app/api/chat/route.ts
import { CostKatana } from 'cost-katana';

export async function POST(req: Request) {
  const katana = new CostKatana({ apiKey: process.env.COST_KATANA_API_KEY });
  const { messages } = await req.json();
  
  const response = await katana.chat.completions.create({
    model: 'gpt-4',
    messages
  });
  
  return Response.json({ response: response.choices[0].message });
}

See: 7-frameworks/nextjs/

FastAPI (Python)

from fastapi import FastAPI
from cost_katana import CostKatana

app = FastAPI()
katana = CostKatana(api_key=os.getenv("COST_KATANA_API_KEY"))

@app.post("/chat")
async def chat(messages: list):
    response = katana.chat.completions.create(
        model="gpt-4",
        messages=messages
    )
    return {"response": response.choices[0].message, "cost": response.cost}

See: 7-frameworks/ for all framework examples

---

Real-World Use Cases

Use Case 1: Multi-Agent Customer Support

Goal: Build a customer support system with multiple AI agents handling different tasks.

Features Used:

• Agents - Create specialized agents
• Workflows - Orchestrate agent interactions
• Memory - Maintain conversation context
• Webhooks - Get notified of escalations

Expected Results:

• 24/7 automated support
• 70% reduction in support costs
• Context-aware responses

---

Use Case 2: Cost-Optimized Content Generation

Goal: Generate marketing content at scale while minimizing costs.

Features Used:

• Cortex Optimization - Compress prompts
• Semantic Caching - Cache similar requests
• Model Routing - Use optimal models
• Cost Tracking - Monitor spending

Expected Results:

• 60-80% cost reduction
• 10x content generation speed
• $10K+ monthly savings

---

Use Case 3: Enterprise AI Platform

Goal: Build a secure, compliant AI platform for enterprise use.

Features Used:

• Key Vault - Secure key management
• Guardrails - Content moderation
• Audit Logs - Compliance tracking
• MFA - Enhanced security
• Team Management - Role-based access

Expected Results:

• SOC 2 compliance ready
• Zero security incidents
• Full audit trail

---

ROI Calculator

Scenario: 1000 AI Requests/Day

Without Cost Katana:

1000 requests/day × $0.045/request = $45/day
Monthly: $1,350
Yearly: $16,425

With Cost Katana (Caching + Optimization):

1000 requests/day × 20% actual calls × $0.03/request = $6/day
Monthly: $180
Yearly: $2,190

SAVINGS: $14,235/year (87% reduction)

ROI Breakdown:

• Semantic Caching: 70-80% hit rate = 70-80% cost savings
• Cortex Optimization: 30-40% token reduction
• Model Routing: 15-25% savings using optimal models
• Failover: Prevent downtime costs

---

Best Practices

1. Security

// ✅ DO: Use environment variables
const apiKey = process.env.COST_KATANA_API_KEY;

// ❌ DON'T: Hardcode API keys
const apiKey = 'ck_live_abc123'; // NEVER DO THIS

2. Error Handling

try {
  const response = await katana.chat.completions.create({...});
} catch (error) {
  if (error.status === 429) {
    // Rate limiting - implement exponential backoff
    await exponentialBackoff();
  } else if (error.status === 500) {
    // Server error - try failover
    await failoverToBackupProvider();
  }
  // Log error for monitoring
  logger.error('AI request failed', { error, context });
}

3. Caching Strategy

// High-precision responses: Higher threshold
const critical = await katana.chat.completions.create({
  model: 'gpt-4',
  messages: [...],
  cache: { enabled: true, threshold: 0.95 } // 95% match required
});

// General queries: Lower threshold
const general = await katana.chat.completions.create({
  model: 'gpt-3.5-turbo',
  messages: [...],
  cache: { enabled: true, threshold: 0.80 } // 80% match OK
});

4. Monitoring

// Always track costs and performance
const response = await katana.chat.completions.create({...});

// Log metrics
logger.info('AI request completed', {
  cost: response.cost,
  tokens: response.usage.total_tokens,
  latency: response.latency,
  cached: response.cached,
  model: response.model
});

---

Repository Structure

costkatana-examples/
├── 1-cost-tracking/          # Track costs across providers
│   ├── http-headers/         # HTTP REST API examples
│   ├── npm-package/          # TypeScript/Node.js examples
│   ├── python-sdk/           # Python SDK examples
│   └── README.md             # Feature documentation
│
├── 10-webhooks/              # Real-time notifications
│   ├── http-headers/         # Webhook API examples
│   ├── npm-package/          # TypeScript webhook handlers
│   ├── python-sdk/           # Python webhook receivers
│   ├── frameworks/           # Express, Next.js, etc.
│   └── README.md             # Webhook guide
│
├── 14-cache/                 # Semantic caching
├── 13-workflows/             # AI orchestration
├── 27-agents/                # AI agents
└── ... (44 feature directories total)

Each feature directory contains:

• README.md - Complete feature documentation
• http-headers/ - Direct HTTP API examples
• npm-package/ - TypeScript/Node.js SDK examples
• python-sdk/ - Python SDK examples
• frameworks/ - Framework-specific integrations (when applicable)

---

Supported Providers

Cost Katana works with all major AI providers:

• ✅ OpenAI (GPT-4, GPT-3.5, DALL-E)
• ✅ Anthropic (Claude 3 Opus/Sonnet/Haiku)
• ✅ AWS Bedrock (All models)
• ✅ Google AI (Gemini, PaLM)
• ✅ Cohere (Command, Embed)
• ✅ Azure OpenAI (All models)
• ✅ DeepSeek (DeepSeek models)
• ✅ Grok (Llama, Mixtral)
• ✅ HuggingFace (Open models)
• ✅ Ollama (Local models)
• ✅ Replicate (All models)

300+ AI models supported across all providers.

---

Getting Help

Documentation

📚 Docs: docs.costkatana.com
🎓 Tutorials: docs.costkatana.com/tutorials
📖 API Reference: docs.costkatana.com/api

Community

💬 Discord: discord.gg/Wcwzw8wM
🐦 Twitter: @costkatana
🐙 GitHub: github.com/costkatana

Support

📧 Email: support@costkatana.com
🌐 Dashboard: costkatana.com/dashboard
📊 Status: status.costkatana.com

---

Contributing

We welcome contributions! Here's how to add your own examples:

1. Fork this repository
2. Create a feature branch (git checkout -b feature/my-example)
3. Add your example following our structure:
   • Include HTTP, TypeScript, and Python versions
   • Add comprehensive README with use cases
   • Include error handling and best practices
   • Test with real API endpoints
4. Commit your changes (git commit -m 'Add: My awesome example')
5. Push to your branch (git push origin feature/my-example)
6. Open a Pull Request

Code Quality Guidelines:

• ✅ Production-ready code with error handling
• ✅ Clear comments explaining the logic
• ✅ Type safety (TypeScript) and type hints (Python)
• ✅ Security best practices (no hardcoded keys)
• ✅ Real-world use cases in README

---

License

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

---

Statistics

📊 Repository Metrics

Total Examples:     300+
Features Covered:   44
Languages:          TypeScript, Python, HTTP
Frameworks:         5 (Express, Next.js, Fastify, NestJS, FastAPI)
Lines of Code:      ~15,000+
Documentation:      50 comprehensive READMEs

Updated:            Weekly
Maintained by:      Cost Katana Team

---

🚀 Ready to Get Started?

Browse Examples • Get API Key • Read Docs • Join Discord

---

Built with ❤️ by the Cost Katana Team

Making AI development simpler, faster, and more cost-effective