> ## Documentation Index
> Fetch the complete documentation index at: https://api.amlyze.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Architecture
AMLYZE works in the background (**asynchronously**). This means that when you send data, you don't have to wait for the full analysis to finish before receiving a response.
### How Requests Work
When you send a request for evaluating a customer or an operation, the system follows these steps:
1. **Validation:** The system checks that the data format is correct.
2. **Creation:** It saves the information and triggers an internal event.
3. **Response:** It sends you a simple `REQUEST_ACCEPTED` or `REQUEST_REJECTED` message.
### Receiving Results (Webhooks)
Because the analysis happens in the background, you must provide an **endpoint (URL)** where AMLYZE can send the results.
* **Flexible Routing:** You can set up multiple endpoints.
* **Smart Filtering:** You can choose to send different types of events (like screening vs. risk scoring) to different URLs.
* **Business Unit Separation:** Each Business Unit can have its own dedicated endpoint to keep data organized.
### Integration Flow
### A Note on Timing
In some rare cases, the analysis result might reach your endpoint **before** you receive the `REQUEST_ACCEPTED` response from the initial request. Your system should be built to handle these results immediately, even if the original task hasn't "officially" finished in your logs yet.