Client Configuration

Python

The Python SDK provides two client types: synchronous (CredoAI) and asynchronous (AsyncCredoAI). Choose based on your application's needs.

TypeScript

The TypeScript SDK provides a single async client created via createCredoAIClient(). All methods return promises.

Creating a Client

Python

from credoai import CredoAI

# Initialize with environment variables (recommended)
client = CredoAI()

# Or with explicit configuration
client = CredoAI(
    api_key="your-api-key",
    base_url="https://api.credo.ai",
    tenant="your-tenant"
)

# Use the client
response = client.use_cases.list()
for uc in response.items:
    print(uc.name)

TypeScript

import { createCredoAIClient } from '@credo-ai/sdk';

// Initialize with environment variables (recommended)
const client = createCredoAIClient('your-tenant');

// Or with explicit configuration
const client = createCredoAIClient('your-tenant', {
  apiKey: 'your-api-key',
  baseURL: 'https://api.credo.ai',
});

// Use the client
const { data, error } = await client.useCases.list();
if (!error) {
  for (const uc of data.items) {
    console.log(uc.name);
  }
}

Async Client (Python)

Python

The AsyncCredoAI client is designed for async applications, web servers, and concurrent operations:

import asyncio
from credoai import AsyncCredoAI

async def main():
    # Use as a context manager (recommended)
    async with AsyncCredoAI() as client:
        response = await client.use_cases.list()
        for uc in response.items:
            print(uc.name)

asyncio.run(main())

Async Context Manager

The async client should be used as a context manager to ensure proper cleanup:

async with AsyncCredoAI() as client:
    # Client is properly initialized
    await client.use_cases.list()
# Client is automatically closed

Manual Lifecycle Management

If you can't use a context manager:

client = AsyncCredoAI()
try:
    await client.use_cases.list()
finally:
    await client.close()

TypeScript

The TypeScript client is async by default — all methods return promises. No context manager or manual cleanup is needed:

import { createCredoAIClient } from '@credo-ai/sdk';

const client = createCredoAIClient('your-tenant');

// All methods are async
const { data, error } = await client.useCases.list();
if (!error) {
  for (const uc of data.items) {
    console.log(uc.name);
  }
}

When to Use Each Client

Use Case	Python	TypeScript
Scripts and CLI tools	CredoAI (sync)	createCredoAIClient()
Jupyter notebooks	CredoAI (sync)	—
FastAPI / async web servers	AsyncCredoAI (async)	—
Express / Node.js servers	—	createCredoAIClient()
Concurrent API calls	AsyncCredoAI (async)	createCredoAIClient()
Django / Flask	CredoAI (sync)	—

Client Options

Python

Option	Type	Description
api_key	str	API key (defaults to CREDOAI_API_KEY env var)
base_url	str	API base URL (defaults to CREDOAI_API_URL env var)
tenant	str	Tenant identifier (defaults to CREDOAI_TENANT env var)

TypeScript

The first argument to createCredoAIClient() is the tenant (falls back to CREDOAI_TENANT env var). The second is an options object:

Option	Type	Description
apiKey	string	API key (defaults to CREDOAI_API_KEY env var)
baseURL	string	API base URL (defaults to CREDOAI_API_URL env var)
token	string	Pre-exchanged bearer token (takes precedence over apiKey)

Accessing Resources

Python

# Synchronous
client = CredoAI()
client.use_cases.list()
client.models.create(data=model_data)
client.vendors.get(vendor_id="...")

# Asynchronous
async with AsyncCredoAI() as client:
    await client.use_cases.list()
    await client.models.create(data=model_data)
    await client.vendors.get(vendor_id="...")

TypeScript

const client = createCredoAIClient('your-tenant');

// All methods are async and return { data, error }
const useCases = await client.useCases.list();
const model = await client.models.create({ name: 'My Model' });
const vendor = await client.vendors.get('vendor-id');

Next Steps

• Follow the quickstart
• Learn about async operations