> ## Documentation Index
> Fetch the complete documentation index at: https://developers.resistant.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Overview

This section explains the supported ways to receive analysis results from the Resistant Documents API.

There are three options:

1. **Polling the result endpoints (default)**
2. **Event notifications via Amazon SQS (optional add-on)**
3. **Webhooks via Svix (optional add-on)**

<Info>
  SQS event notifications and Svix webhooks are available subject to prior agreement and may involve an additional fee.
</Info>

***

## Option 1 — Polling (default)

By default, customers fetch results by polling the relevant result endpoints (e.g., fraud/quality/classification/decision) until processing completes.

**Recommended polling strategy**

* Implement **exponential backoff**
* Cap the polling interval at **45 seconds**
* Stop polling after **15 minutes** (hard analysis timeout)

**Best for**

* Most integrations
* Teams who want the simplest implementation with no additional infrastructure

➡️ Next: [Polling for results](/receiving-results/polling)

***

## Option 2 — Amazon SQS event notifications

Instead of polling, you can subscribe to events delivered to your **Amazon SQS** queue. This allows your application to react to results asynchronously and scale via event-driven processing.

**How it works (high level)**

* Events are delivered to your SQS queue.
* Resistant AI assumes a role in your AWS account using STS; the `tenant_id` is used as the `ExternalId`.
* Each event includes a `result_url` you can use to fetch full results via the API.

**Supported event types**

* `documents.analysis.fraud.finished`
* `documents.analysis.quality.finished`
* `documents.analysis.classification.finished`

**Notable limitations**

* One event destination per tenant.
* Publishing does not specify a KMS key in the API call; your SQS queue policy must support your encryption setup.

**Best for**

* AWS-native customers
* High-throughput integrations where polling load is undesirable

➡️ Next: [Amazon SQS event notifications](/receiving-results/amazon-sqs)

***

## Option 3 — Webhooks via Svix

Webhooks are another alternative to polling. Resistant AI can send HTTPS webhook requests to your endpoint using **Svix** as the delivery provider.

**How it works (high level)**

* You provide a callback URL.
* Svix delivers events to your endpoint.
* Webhooks include signature headers for verification.
* On delivery failures, Svix retries using exponential backoff.

**Event types include**

* `documents.analysis.fraud.finished`
* `documents.adaptive_decision.finished`
* `documents.quality.finished`
* `documents.classification.finished`

**Best for**

* Customers who prefer HTTP callbacks over AWS SQS
* Integrations that already operate webhook receivers

➡️ Next: [Webhooks (Svix)](/receiving-results/webhooks)

***

## Choosing the right option

| Feature               | Polling                              | SQS                   | Webhooks            |
| --------------------- | ------------------------------------ | --------------------- | ------------------- |
| Infrastructure needed | None                                 | AWS account           | HTTPS endpoint      |
| Latency               | Variable (depends on poll frequency) | Near real-time        | Near real-time      |
| Reliability           | Depends on client                    | High (SQS durability) | High (Svix retries) |
| AWS-native            | No                                   | Yes                   | No                  |
| Setup complexity      | Low                                  | Medium                | Low–Medium          |

***