Sandbox & Testing Environment
Last Updated: 2026-01-24
The Abba Baba platform provides a powerful Sandbox Environment for every developer. This is a safe, isolated space where you can build, test, and debug your AI agent without any risk of affecting live data or incurring real-world costs.
Sandbox Features & Limitations
The Sandbox is designed to mirror the production environment's functionality with a few key differences:
- Separate API Keys: Sandbox keys are distinct from your production keys. They will typically be prefixed with
aba_test_.... - Mock Data Catalog: The sandbox uses a static but representative mock catalog of products. It does not connect to live merchant data. This data set includes products with varying data quality and specifications to help you test your agent's parsing and decision logic.
- No Real-World Effects: API calls made with a sandbox key will not trigger real commissions, create real orders, or affect your production analytics.
- Simulated Events: The sandbox environment can simulate events like price changes or out-of-stock notifications, allowing you to test your agent's reactivity.
Getting Your Sandbox API Key
Every developer account has access to the sandbox.
- Navigate to your Developer Dashboard (opens in a new tab).
- Go to the "Agents" section.
- When you "Create New Agent," you will have the option to generate either a Production Key or a Sandbox Key.
- Generate and securely store your new Sandbox API key. You can have multiple sandbox agents for different testing purposes.
Using the Sandbox
The Sandbox uses the same API endpoints as our production environment. The only difference is the API key you provide in the Authorization header.
Switching Between Environments
A best practice is to store both your production and sandbox keys as environment variables and switch between them based on your application's environment.
import os
# Set an environment variable like `APP_ENV=production` or `APP_ENV=development`
APP_ENVIRONMENT = os.getenv("APP_ENV", "development")
if APP_ENVIRONMENT == "production":
API_KEY = os.getenv("ABA_PROD_API_KEY")
else:
API_KEY = os.getenv("ABA_SANDBOX_API_KEY")
headers = {
"Authorization": f"Bearer {API_KEY}",
"Content-Type": "application/json"
}
# Your API call logic remains the same
# ...Testing Commission & Referrals in the Sandbox
You can test your agent's full referral and commission-earning logic in the sandbox.
- Mock Checkout:
checkoutUrllinks generated by a sandbox agent will point to a mock checkout page. - Simulated Purchases: "Purchasing" an item on this mock page will simulate a successful referral event.
- Sandbox Analytics: This simulated purchase will generate corresponding data in your sandbox analytics dashboard, allowing you to verify that your agent's commission tracking and performance monitoring logic is working correctly before deploying to production.