Skip to main content

​
Client Configuration

The BaytClient class provides flexible configuration options for timeouts, retries, connection pooling, and more.

​
Basic Initialization

from baytos.claro import BaytClient

client = BaytClient(api_key="your_api_key")

​
Configuration Options

​
API Key

The API key is required and can be provided in several ways: 
import os
from baytos.claro import BaytClient

# Method 1: Environment variable (recommended)
client = BaytClient(api_key=os.getenv("BAYT_API_KEY"))

# Method 2: Direct (not recommended for production)
client = BaytClient(api_key="your_api_key_here")

# Method 3: From config file
import json
with open('config.json') as f:
    config = json.load(f)
client = BaytClient(api_key=config['bayt_api_key'])

​
Base URL

The SDK uses https://api.baytos.ai by default: 
# Production (default)
client = BaytClient(api_key="...")
# Uses: https://api.baytos.ai
Enterprise Self-Hosting: If you’re running a self-hosted instance of Bayt OS (available for enterprise customers), you can override the base_url parameter to point to your instance:
client = BaytClient(
    api_key="...",
    base_url="https://bayt.yourcompany.com"
)
Contact support@baytos.ai to learn more about self-hosting options.

​
Retries and Timeouts

Configure retry behavior for resilient applications: 
client = BaytClient(
    api_key="...",
    max_retries=3,        # Number of retry attempts (default: 3)
    timeout=30,           # Request timeout in seconds (default: 30)
    retry_delay=1.0       # Initial retry delay in seconds (default: 1.0)
)
The SDK uses exponential backoff: retry_delay * (2 ** attempt). With retry_delay=1.0, retries occur at 1s, 2s, 4s, etc.

​
Retry Behavior

The SDK automatically retries on:
  • 429 (Rate Limit) - Respects Retry-After header
  • 5xx (Server Errors) - Uses exponential backoff
  • Network Errors - Connection timeouts, DNS failures
The SDK does NOT retry on:
  • 4xx (Client Errors) - Except 429
  • Authentication Errors (401, 403)
  • Not Found (404)
# Aggressive retries for unreliable networks
client = BaytClient(
    api_key="...",
    max_retries=5,
    timeout=60
)

# Disable retries (fast-fail)
client = BaytClient(
    api_key="...",
    max_retries=0
)

​
Connection Pooling

The SDK uses requests.Session for automatic connection pooling: 
client = BaytClient(api_key="...")

# Connection is established and pooled
prompt1 = client.get_prompt("@workspace/p1:v1")

# Connection is reused (faster)
prompt2 = client.get_prompt("@workspace/p2:v1")
prompt3 = client.get_prompt("@workspace/p3:v1")
Benefits:
  • Faster subsequent requests (no TCP handshake)
  • Lower latency
  • Reduced server load
  • Automatic keep-alive management
Reuse the same BaytClient instance across your application for optimal performance.

​
Custom Headers

Add custom headers to all requests: 
client = BaytClient(
    api_key="...",
    headers={
        "X-App-Version": "1.2.3",
        "X-Environment": "production"
    }
)

​
Proxy Configuration

Configure HTTP/HTTPS proxies: 
client = BaytClient(
    api_key="...",
    proxies={
        "http": "http://proxy.example.com:8080",
        "https": "https://proxy.example.com:8443"
    }
)

# With authentication
client = BaytClient(
    api_key="...",
    proxies={
        "http": "http://user:pass@proxy.example.com:8080",
        "https": "https://user:pass@proxy.example.com:8443"
    }
)

​
SSL/TLS Configuration

​
Verify SSL Certificates

# Enable SSL verification (default)
client = BaytClient(api_key="...", verify_ssl=True)

# Disable SSL verification (not recommended)
client = BaytClient(api_key="...", verify_ssl=False)

# Use custom CA bundle
client = BaytClient(
    api_key="...",
    verify_ssl="/path/to/ca-bundle.crt"
)
Disabling SSL verification exposes you to man-in-the-middle attacks. Only use in controlled development environments.

​
User Agent

Customize the User-Agent header: 
client = BaytClient(
    api_key="...",
    user_agent="MyApp/1.0 (contact@example.com)"
)

​
Production Configuration Example

Complete production-ready configuration: 
import os
from baytos.claro import BaytClient

client = BaytClient(
    # Authentication
    api_key=os.getenv("BAYT_API_KEY"),

    # Performance
    max_retries=3,
    timeout=30,
    retry_delay=1.0,

    # Monitoring
    headers={
        "X-App-Version": os.getenv("APP_VERSION", "unknown"),
        "X-Environment": os.getenv("ENVIRONMENT", "production"),
        "X-Instance-ID": os.getenv("INSTANCE_ID", "unknown")
    },

    # Security
    verify_ssl=True
)

​
Configuration Best Practices

Create one client instance and reuse it across your application:
# ✅ Good: Single instance
client = BaytClient(api_key=os.getenv("BAYT_API_KEY"))

def get_support_prompt():
    return client.get_prompt("@workspace/support:v1")

def get_marketing_prompt():
    return client.get_prompt("@workspace/marketing:v1")

# ❌ Bad: Multiple instances
def get_support_prompt():
    client = BaytClient(api_key=os.getenv("BAYT_API_KEY"))
    return client.get_prompt("@workspace/support:v1")
Choose timeouts based on your use case:
  • Interactive applications: 5-10 seconds
  • Background jobs: 30-60 seconds
  • Batch processing: 60-120 seconds
# Interactive
client = BaytClient(api_key="...", timeout=10)

# Background processing
client = BaytClient(api_key="...", timeout=60)
Use different configurations per environment:
import os

ENV = os.getenv("ENVIRONMENT", "production")

if ENV == "development":
    client = BaytClient(
        api_key=os.getenv("BAYT_API_KEY"),
        max_retries=0,  # Fast-fail in dev
        timeout=10
    )
else:  # production
    client = BaytClient(
        api_key=os.getenv("BAYT_API_KEY"),
        max_retries=3,
        timeout=30
    )

​
Next Steps

Working with Prompts

Learn how to fetch and use prompts

Error Handling

Handle errors and retries gracefully

Advanced Features

Explore rate limiting, pooling, and performance optimization

API Reference

Complete API documentation