Amazon Rekognition endpoints and quotas - AWS General Reference

Amazon Rekognition endpoints and quotas

The following are the service endpoints and service quotas for this service. To connect programmatically to an AWS service, you use an endpoint. In addition to the standard AWS endpoints, some AWS services offer FIPS endpoints in selected Regions. For more information, see AWS service endpoints. Service quotas, also referred to as limits, are the maximum number of service resources or operations for your AWS account. For more information, see AWS service quotas.

Service Endpoints

Region Name Region Endpoint Protocol
US East (Ohio) us-east-2

rekognition.us-east-2.amazonaws.com

rekognition-fips.us-east-2.amazonaws.com

HTTPS

HTTPS

US East (N. Virginia) us-east-1

rekognition.us-east-1.amazonaws.com

rekognition-fips.us-east-1.amazonaws.com

HTTPS

HTTPS

US West (N. California) us-west-1

rekognition.us-west-1.amazonaws.com

rekognition-fips.us-west-1.amazonaws.com

HTTPS

HTTPS

US West (Oregon) us-west-2

rekognition.us-west-2.amazonaws.com

rekognition-fips.us-west-2.amazonaws.com

HTTPS

HTTPS

Asia Pacific (Mumbai) ap-south-1 rekognition.ap-south-1.amazonaws.com HTTPS
Asia Pacific (Seoul) ap-northeast-2 rekognition.ap-northeast-2.amazonaws.com HTTPS
Asia Pacific (Singapore) ap-southeast-1 rekognition.ap-southeast-1.amazonaws.com HTTPS
Asia Pacific (Sydney) ap-southeast-2 rekognition.ap-southeast-2.amazonaws.com HTTPS
Asia Pacific (Tokyo) ap-northeast-1 rekognition.ap-northeast-1.amazonaws.com HTTPS
Canada (Central) ca-central-1

rekognition.ca-central-1.amazonaws.com

rekognition-fips.ca-central-1.amazonaws.com

HTTPS

HTTPS

Europe (Frankfurt) eu-central-1 rekognition.eu-central-1.amazonaws.com HTTPS
Europe (Ireland) eu-west-1 rekognition.eu-west-1.amazonaws.com HTTPS
Europe (London) eu-west-2 rekognition.eu-west-2.amazonaws.com HTTPS
AWS GovCloud (US-West) us-gov-west-1

rekognition.us-gov-west-1.amazonaws.com

rekognition-fips.us-gov-west-1.amazonaws.com

HTTPS

HTTPS

The following are differences for certain Amazon Rekognition features and AWS Regions.

Amazon Rekognition Video streaming API

The Amazon Rekognition Video streaming API is available in the following regions only.

  • US East (N. Virginia)

  • US West (Oregon)

  • Asia Pacific (Tokyo)

  • Europe (Frankfurt)

  • Europe (Ireland)

Amazon Rekognition Custom Labels

Amazon Rekognition Custom Labels is available in the following regions only.

  • US East (N. Virginia)

  • US East (Ohio)

  • US West (Oregon)

  • Europe (Ireland)

  • Europe (London)

  • Europe (Frankfurt)

  • Asia Pacific (Mumbai)

  • Asia Pacific (Singapore)

  • Asia Pacific (Sydney)

  • Asia Pacific (Tokyo)

  • Asia Pacific (Seoul)

Canada (Central) Region

The Canada (Central) Region supports the following operations only.

Service Quotas

The quotas listed on this page are defaults. You can request a quota increase for Amazon Rekognition using the AWS Support Center. To request a quota increase for a Amazon Rekognition Transactions Per Second (TPS) limit, follow the instructions at Default Quotas in Amazon Rekognition.

Note

These limits may be different in different regions. Making a case to change a limit affects the API operation you request, in the region you request it. Other API operations and regions are not affected.

Resource Default

Transactions per second per account for individual Amazon Rekognition Image data plane operations:

  • US East (Ohio) Region – 5

  • US East (N. Virginia) Region – 50

  • US West (N. California) Region – 5

  • US West (Oregon) Region – 50

  • Asia Pacific (Mumbai) Region – 5

  • Asia Pacific (Seoul) Region – 5

  • Asia Pacific (Singapore) Region – 5

  • Asia Pacific (Sydney) Region – 5

  • Asia Pacific (Tokyo) Region – 5

  • Canada (Central) – 5 (For supported operations, see Service Endpoints).

  • Europe (Frankfurt) Region – 5

  • Europe (Ireland) Region – 50

  • Europe (London) Region – 5

  • AWS GovCloud (US-West) – 5

Transactions per second per account for the personal protective equipment data plane operation:

In each Region that Amazon Rekognition Image supports – 5

Transactions per second per account for individual Amazon Rekognition Image control plane operations:

In each Region that Amazon Rekognition Image supports – 5

Transactions per second per account for individual stored video start operations:

In each Region that Amazon Rekognition Video supports – 5

Note that StartCelebrityRecognition is not available in AWS GovCloud.

Transactions per second per account for individual Amazon Rekognition Video stored video get operations:

  • US East (Ohio) Region – 5

  • US East (N. Virginia) Region – 20

  • US West (N. California) Region – 5

  • US West (Oregon) Region – 20

  • Asia Pacific (Mumbai) Region – 5

  • Asia Pacific (Seoul) Region – 5

  • Asia Pacific (Singapore) Region – 5

  • Asia Pacific (Sydney) Region – 5

  • Asia Pacific (Tokyo) Region – 5

  • Europe (Frankfurt) Region – 5

  • Europe (Ireland) Region – 20

  • Europe (London) Region – 5

  • AWS GovCloud (US-West) –20 (Note that GetCelebrityRecognition is not available in this region.)

Maximum number of concurrent stored video jobs per account 20
Maximum number of streaming video stream processors per account that can simultaneously exist In each Region that Amazon Rekognition Video supports – 10

Transactions per second per account for individual streaming video operations:

In each Region that Amazon Rekognition Video supports – 1

Transactions per second per account for individual Amazon Rekognition Custom Labels control plane operations:

In each Region that Amazon Rekognition Custom Labels supports – 5
Maximum number of Amazon Rekognition Custom Labels projects per account. 100
Maximum number of Amazon Rekognition Custom Labels models per project. 100
Maximum number of concurrent Amazon Rekognition Custom Labels training jobs per account.
  • All regions except Asia Pacific (Sydney) – 2

  • Asia Pacific (Sydney) – 1

Maximum number of concurrently running Amazon Rekognition Custom Labels models per account. 2
Maximum inference units per started model. 5
Maximum number of images per dataset. 250,000

For more information, see Amazon Rekognition Quotas.