# Amazon Bedrock

Before configuring Amazon Bedrock as your AI provider, ensure you have the following:

* An [AWS account](https://signin.aws.amazon.com/signup?request_type=register).

* An [IAM User](https://docs.aws.amazon.com/IAM/latest/UserGuide/id_users_create.html) with the necessary permissions for Amazon Bedrock.

* Access to the required models. In commercial AWS regions, access to foundation models is enabled by default for accounts with the necessary AWS Marketplace permissions. For Anthropic models, you need to submit a one-time use case form to start using them. For details, see [Request access to models](https://docs.aws.amazon.com/bedrock/latest/userguide/model-access.html).

Once the prerequisites are ready, obtain an access key and configure the provider.

Procedure: Get an access key

1. Access the [IAM console](https://console.aws.amazon.com/iam/), find your user and review the Permissions policies section under the Permissions tab.

2. Add a new inline policy for the Bedrock service.

3. Configure the new inline policy to allow the following actions: `InvokeModel`, `InvokeModelWithResponseStream`, `ListFoundationModels`, `GetFoundationModel`, and `ListInferenceProfiles`.

4. In the Security credentials tab, generate an [access key](https://docs.aws.amazon.com/IAM/latest/UserGuide/id_credentials_access-keys.html) for your user. When creating an access key, specify Third-party service as a use case.

5. Save the generated access key ID and secret. You will need them when configuring the provider in JetBrains Central Console.

Procedure: Add Amazon Bedrock provider

1. In the sidebar, under AI governance, select Providers.

If you don't see this page, or the controls on this page are inactive, your role doesn't have the necessary [permissions](roles.html#permissions).

2. Click Add provider and select Amazon Bedrock. The Amazon Bedrock provider card appears on the page.

3. In the Region field, specify an [AWS region](https://docs.aws.amazon.com/general/latest/gr/bedrock.html) that supports Amazon Bedrock. Ensure it matches the region where the required models are available to your account.

4. Select the Access key tab and enter the [access key ID and access key secret](#get-access-key).

5. Click Connect.

> **Tip:**
> If the connection fails, ensure the access key ID and access key secret are correct, the IAM user has the required permissions, and the specified region matches the region where the required models are available to your account.

Once the connection is established, the available models appear in the provider card.

To add other models for this provider, click Add model in the Models table. When you enter the Provider model ID, use the [exact ID that Amazon Bedrock uses for the model
in your region](#amazon-bedrock-model-id). For the generic procedure, see [Add a model](ai-providers-manage-models.html#add-model).

To edit the provider configuration, click ![Ellipsis icon](images/union.svg) in the provider card, then select Edit credentials and update the necessary fields.

## Amazon Bedrock model ID

When you add or edit a model on the Amazon Bedrock provider card, the Provider model ID must match an exact ID that Amazon Bedrock uses for that model. Which ID is valid depends on the AWS region set in the Region field in the provider card and on the inference type you choose.

JetBrains Central Console pre-fills the Provider model ID with a default value. For most models, you still need to adjust it. A single model can have up to three IDs, one per inference type:

* A base model ID without a prefix, for example, `anthropic.claude-opus-4-8`. Use it for In-Region inference, where the request stays in the Region specified in the provider card.

* A geographic inference profile ID with a regional prefix such as `us`, `eu`, `jp`, or `au`, for example, `us.anthropic.claude-opus-4-8`. Use it for cross-region routing within that geography.

* A global inference profile ID with the `global` prefix, for example, `global.anthropic.claude-opus-4-8`. Use it for cross-region routing worldwide.

The available inference types depend on the model and region, and a single model can support one or more of them in a given region. When several types are available, choose one based on your data residency requirements: In-Region for a single region, Geo for data residency within a geography, or Global for no residency constraints. To check which types each model supports per region, see [Regional availability](https://docs.aws.amazon.com/bedrock/latest/userguide/models-region-compatibility.html).

Once you know which inference type you want, copy the exact ID from the model's card.

Procedure: Get a model ID

1. Open [Models at a glance](https://docs.aws.amazon.com/bedrock/latest/userguide/model-cards.html), then open the model's card.

2. In the Regional Availability section, confirm that the region from the provider card supports your chosen inference type (In-Region, Geo, or Global).

3. In the Programmatic Access section, copy the ID for a supported type: the base Model ID, a Geo inference ID, or the Global inference ID.

You now have the exact ID that Amazon Bedrock expects for your region and inference type.

> **Tip:**
> The same IDs are available in the [Amazon Bedrock console](https://console.aws.amazon.com/bedrock), where you can confirm which IDs are enabled for your account and region.

Paste the copied ID into the [Provider model ID](ai-providers-manage-models.html) field in the provider card. JetBrains Central Console sends this ID to Amazon Bedrock unchanged.

Amazon Bedrock validates the model ID when the [connectivity check](ai-providers-manage-models.html#check-model-connectivity) runs. If the ID isn't valid for the model in your region (for example, you used a base model ID in a region where the model supports only cross-region routing), the check fails.