# Migrate from AWS Rekognition to SightRadar

> Keep your boto3 / AWS SDK code. Change two lines — the endpoint and the credentials — and your existing face-recognition calls run on SightRadar at roughly half the per-photo price, with no AWS account or quota tickets.

Source: https://sightradar.com/migrate

## Why it's drop-in

SightRadar uses the same request and response shapes as AWS Rekognition for the face operations, so an existing integration keeps working. Supported operations:

- CreateCollection
- IndexFaces
- SearchFacesByImage
- SearchFaces
- CompareFaces
- DetectFaces
- ListFaces / DeleteFaces
- ListCollections / DeleteCollection

## What changes

- Point the endpoint at `https://api.sightradar.com` instead of the AWS Rekognition regional endpoint.
- Use a SightRadar API key (`frs_<prefix>_<secret>`) sent as `Authorization: Bearer <key>` — no IAM users, roles, or policies.
- The official Python shim (`sightradar-rekognition-shim`) maps Rekognition method names so you can keep your boto3-style calls.

## What you gain

- Roughly half the per-photo cost ($0.00093 real-time, $0.00062 batch vs $0.00125 Group-1).
- No default TPS quotas to raise by support ticket, no per-operation caps.
- A real web console to browse collections, run searches, and see per-operation cost.
- Wider image support: WebP and HEIC/HEIF decode natively (Rekognition is JPEG/PNG only).

Side-by-side before/after and the method mapping: https://sightradar.com/migrate