Amazon Bedrock
The Amazon Bedrock provider for the AI SDK contains language model support for the Amazon Bedrock APIs.
Start with AI Gateway for models in its catalog (recommended), or use @ai-sdk/amazon-bedrock to connect directly to Amazon Bedrock.
Using AI Gateway
The AI SDK uses AI Gateway by default for model strings. Gateway provides access to models across providers through a single API key, with usage and cost tracking. Install ai:
$ pnpm add ai
$ npm install ai
$ yarn add ai
$ bun add ai
Set an AI Gateway API key in your environment:
Vercel deployments can also use OIDC authentication. See the AI Gateway provider documentation for configuration and the model catalog for available models.
The example sets providerOptions.gateway.order to ['bedrock']. This sends the request to Amazon Bedrock first, then falls back to other available providers of the same model if Amazon Bedrock is unavailable or the request fails. order sets a preference; it does not restrict fallback to Amazon Bedrock. See provider ordering.
The dropdown lists models served by Amazon Bedrock through Gateway with a corresponding direct-provider example. Provider availability depends on the model and your team's Provider Allowlist.
Using @ai-sdk/amazon-bedrock
For direct access to Amazon Bedrock, install ai and @ai-sdk/amazon-bedrock:
$ pnpm add ai @ai-sdk/amazon-bedrock
$ npm install ai @ai-sdk/amazon-bedrock
$ yarn add ai @ai-sdk/amazon-bedrock
$ bun add ai @ai-sdk/amazon-bedrock
Set your Amazon Bedrock API credential in the AWS_BEARER_TOKEN_BEDROCK environment variable:
Use the direct provider for AWS authentication, account-specific inference profiles, or Bedrock-specific APIs. Gateway and Bedrock use different model IDs.
The following sections document the direct @ai-sdk/amazon-bedrock provider,
including its model factories and settings. Gateway model and feature
availability can differ.
Prerequisites
Access to Amazon Bedrock foundation models isn't granted by default. In order to gain access to a foundation model, an IAM user with sufficient permissions needs to request access to it through the console. Once access is provided to a model, it is available for all users in the account.
See the Model Access Docs for more information.
Authentication
The Amazon Bedrock provider supports two authentication methods with automatic fallback: API key authentication (recommended) and AWS SigV4 authentication using IAM credentials or the AWS SDK credentials chain.
Using an API Key (Recommended)
API key authentication provides a simpler setup compared to AWS SigV4 authentication. When an API key is supplied, it takes precedence over any SigV4 credentials.
Generate a bedrock API key from the AWS console, then either set it as the AWS_BEARER_TOKEN_BEDROCK environment variable:
or pass it directly to createAmazonBedrock:
When apiKey is omitted, the provider falls back to AWS_BEARER_TOKEN_BEDROCK, and then to SigV4 authentication if neither is set.
Using IAM Access Key and Secret Key
Step 1: Creating AWS Access Key and Secret Key
To get started, you'll need to create an AWS access key and secret key. Here's how:
Login to AWS Management Console
- Go to the AWS Management Console and log in with your AWS account credentials.
Create an IAM User
- Navigate to the IAM dashboard and click on "Users" in the left-hand navigation menu.
- Click on "Create user" and fill in the required details to create a new IAM user.
- Make sure to select "Programmatic access" as the access type.
- The user account needs the
AmazonBedrockFullAccesspolicy attached to it.
Create Access Key
- Click on the "Security credentials" tab and then click on "Create access key".
- Click "Create access key" to generate a new access key pair.
- Download the
.csvfile containing the access key ID and secret access key.
Step 2: Configuring the Access Key and Secret Key
Within your project add a .env file if you don't already have one. This file will be used to set the access key and secret key as environment variables. Add the following lines to the .env file:
Remember to replace YOUR_ACCESS_KEY_ID, YOUR_SECRET_ACCESS_KEY, and YOUR_REGION with the actual values from your AWS account.
Using AWS SDK Credentials Chain (instance profiles, instance roles, ECS roles, EKS Service Accounts, etc.)
When using AWS SDK, the SDK will automatically use the credentials chain to determine the credentials to use. This includes instance profiles, instance roles, ECS roles, EKS Service Accounts, etc. A similar behavior is possible using the AI SDK by not specifying the accessKeyId and secretAccessKey, sessionToken properties in the provider settings and instead passing a credentialProvider property.
Usage:
@aws-sdk/credential-providers package provides a set of credential providers that can be used to create a credential provider chain.
$ pnpm add @aws-sdk/credential-providers
$ npm install @aws-sdk/credential-providers
$ yarn add @aws-sdk/credential-providers
$ bun add @aws-sdk/credential-providers
Provider Instance
You can import the default provider instance amazonBedrock from @ai-sdk/amazon-bedrock:
If you need a customized setup, you can import createAmazonBedrock from @ai-sdk/amazon-bedrock and create a provider instance with your settings:
Omitted options use the environment variables below. When both
accessKeyId and secretAccessKey are strings, SigV4 uses a sessionToken
only if you pass one - not AWS_SESSION_TOKEN from the environment - so
static keys are not mixed with workload tokens (e.g. EKS IRSA).
You can use the following optional settings to customize the Amazon Bedrock provider instance:
-
region string
The AWS region that you want to use for the API calls. It uses the
AWS_REGIONenvironment variable by default. -
accessKeyId string
The AWS access key ID that you want to use for the API calls. It uses the
AWS_ACCESS_KEY_IDenvironment variable by default. -
secretAccessKey string
The AWS secret access key that you want to use for the API calls. It uses the
AWS_SECRET_ACCESS_KEYenvironment variable by default. -
sessionToken string
Optional. For temporary credentials. With both access keys set as strings, pass the token here if needed;
AWS_SESSION_TOKENfrom the environment is not used. If either key is taken from the environment, omittingsessionTokenallowsAWS_SESSION_TOKEN. -
credentialProvider () => Promise<{ accessKeyId: string; secretAccessKey: string; sessionToken?: string; }>
Optional. The AWS credential provider chain that you want to use for the API calls. It uses the specified credentials by default.
-
apiKey string | PromiseLike<string> | (() => string | PromiseLike<string>)
Pass a callback to resolve the API key before each request. The callback can return a string or a promise; token caching and renewal are handled by the caller. A callback or promise selects API key authentication and must resolve to a non-empty string; lookup failures do not switch authentication methods.
Optional. API key for authenticating requests using Bearer token authentication. When provided, this will be used instead of AWS SigV4 authentication. It uses the
AWS_BEARER_TOKEN_BEDROCKenvironment variable by default. -
baseURL string
Optional. Base URL for the Bedrock API calls. Useful for custom endpoints or proxy configurations. When omitted, the provider also supports the service-specific
AWS_ENDPOINT_URL_BEDROCK_RUNTIMEandAWS_ENDPOINT_URL_BEDROCK_AGENT_RUNTIMEenvironment variables, which take precedence over the globalAWS_ENDPOINT_URLenvironment variable. -
headers Record<string, string>
Optional. Custom headers to include in the requests.
-
fetch (input: RequestInfo, init?: RequestInit) => Promise<Response>
Optional. Custom fetch implementation. You can use it as a middleware to intercept requests, or to provide a custom fetch implementation for e.g. testing.
Language Models
You can create models that call the Bedrock API using the provider instance.
The first argument is the model id, e.g. meta.llama3-70b-instruct-v1:0.
Amazon Bedrock models also support some model specific provider options that are not part of the standard call settings.
You can pass them in the providerOptions argument:
Documentation for additional settings based on the selected model can be found within the Amazon Bedrock Inference Parameter Documentation.
You can use Amazon Bedrock language models to generate text with the generateText function:
Amazon Bedrock language models can also be used in the streamText function
(see AI SDK Core).
Structured Outputs
For Anthropic models, you can use the structuredOutputMode provider option to
choose how Amazon Bedrock generates structured outputs:
"outputFormat": Use Bedrock's nativeoutput_config.formatparameter."jsonTool": Use a specialjsontool with required tool choice."auto": Use native structured output when supported and otherwise fall back to the JSON tool. This is the default.
In "auto" mode, supported Claude 5 models use native output for compatible
schemas. Recursive references, open objects, and schemas with more than ten
optional properties fall back to the JSON tool, or JSON instructions for models
that reject forced tool use.
The JSON-tool fallback limits which application tools are available:
toolChoice: 'none' sends only the internal JSON response tool, and a named tool
choice sends only that application tool alongside the JSON response tool. The
fallback can produce a JSON response without calling the named application tool;
verify the selected call if your application requires it.
In "auto" mode, Claude Sonnet 4.6 and Claude Haiku 4.5 use the JSON tool
fallback because native structured output can be unreliable across workloads
and Bedrock accounts. You can set "outputFormat" to opt into native structured
output when it works for your account and workload, or use "jsonTool" to force
the fallback for any model:
The legacy providerOptions.bedrock key remains supported. For parity with the
Anthropic provider, providerOptions.anthropic.structuredOutputMode is also
honored for Anthropic models on Bedrock. When both are provided, the
amazonBedrock or legacy bedrock value takes precedence.
For an Anthropic application inference profile, prefer the Bedrock Anthropic provider. It uses Bedrock's native InvokeModel API, which accepts application inference profile ARNs, and identifies the model as Anthropic automatically:
Use modelFamily: 'anthropic' only when you need to use the generic
@ai-sdk/amazon-bedrock provider, which uses Bedrock's Converse API (for
example, when IAM permits Converse but not InvokeModel). Application inference
profile ARNs do not identify their underlying model, so declaring the family
allows Anthropic-specific features, including native structured output:
For an underlying model that does not reliably support native structured
output, set structuredOutputMode to "jsonTool".
File Inputs
Amazon Bedrock supports file inputs in combination with specific models, e.g.
anthropic.claude-3-haiku-20240307-v1:0.
The Amazon Bedrock provider supports file inputs, e.g. PDF files.
Guardrails
You can use the bedrock provider options to utilize Amazon Bedrock Guardrails:
Tracing information will be returned in the provider metadata if you have tracing enabled.
See the Amazon Bedrock Guardrails documentation for more information.
Guard Content
You can mark individual text or image parts as guard content using providerOptions on the part itself. This wraps the part in a guardContent block, which tells Bedrock Guardrails to evaluate that content specifically.
For text parts, you can also specify guardContentQualifiers to indicate how the content should be treated by the guardrail (e.g. as a grounding source, a query, or guard content):
Image parts can also be marked as guard content:
The available guardContentQualifiers are:
'grounding_source'— content used as a grounding source for the guardrail'query'— content treated as a query to evaluate'guard_content'— content to be evaluated by the guardrail
See the Amazon Bedrock Guardrails with Converse API documentation for more information.
Citations
Amazon Bedrock supports citations for document-based inputs across compatible models. When enabled:
- Some models can read documents with visual understanding, not just extracting text
- Models can cite specific parts of documents you provide, making it easier to trace information back to its source (the AI SDK does not return citation data in results yet)
Cache Points
Amazon Bedrock prompt caching is currently in preview release. To request access, visit the Amazon Bedrock prompt caching page.
In messages, you can use the providerOptions property to set cache points. Set the bedrock property in the providerOptions object to { cachePoint: { type: 'default' } } to create a cache point.
You can also specify a TTL (time-to-live) for cache points using the ttl property. Supported values are '5m' (5 minutes, default) and '1h' (1 hour). The 1-hour TTL is only supported by Claude Opus 4.5, Claude Haiku 4.5, and Claude Sonnet 4.5.
When using multiple cache points with different TTLs, cache entries with longer TTL must appear before shorter TTLs (i.e., 1-hour cache entries must come before 5-minute cache entries).
Cache usage information is returned in the providerMetadata object. See examples below.
Cache points have model-specific token minimums and limits. For example, Claude 3.5 Sonnet v2 requires at least 1,024 tokens for a cache point and allows up to 4 cache points. See the Amazon Bedrock prompt caching documentation for details on supported models, regions, and limits.
Cache points also work with streaming responses:
Provider Metadata
The following Bedrock-specific metadata may be returned in providerMetadata.bedrock:
- trace (optional) Guardrail tracing information (when tracing is enabled).
- performanceConfig (optional)
Performance configuration, e.g.
{ latency: 'optimized' }. - serviceTier (optional)
Service tier information, e.g.
{ type: 'on-demand' }. - usage (optional)
Cache token usage details including
cacheWriteInputTokensandcacheDetails. - stopSequence string | null The stop sequence that triggered the stop, if any.
Reasoning
Amazon Bedrock supports model creator-specific reasoning features:
- Anthropic (e.g.
claude-sonnet-4-5-20250929): enable via thereasoningConfigprovider option and specifying a thinking budget in tokens (minimum:1024, maximum:64000). - Amazon (e.g.
us.amazon.nova-2-lite-v1:0): enable via thereasoningConfigprovider option and specifying a maximum reasoning effort level ('low' | 'medium' | 'high').
You can also use the top-level reasoning parameter.
Bedrock combines partial providerOptions.amazonBedrock.reasoningConfig with
portable reasoning, preserving explicit values and deriving missing fields for
the model's reasoning API. An explicit maxReasoningEffort replaces the effort
mapping without emitting a coercion warning for the overridden value. For
budget-based Anthropic models, { budgetTokens: 5000 } preserves that budget and
uses portable reasoning to fill a missing type: 'enabled'. Type-only,
display-only, and empty configurations also retain portable reasoning.
Effort-based Anthropic models can combine portable effort with an explicit
thinking configuration. A disabled type prevents effort/budget derivation;
portable reasoning: 'none' requests disabled thinking on Anthropic models even
when an enabled type or budget is supplied. Omitting portable reasoning or using
'provider-default' uses the provider configuration. The legacy
bedrock provider options key follows the same rules. See the
reasoning guide for precedence details.
Both reasoning: 'none' and reasoningConfig: { type: 'disabled' } explicitly
send thinking: { type: 'disabled' } for Anthropic models that support it.
For Claude 3.5 and earlier, the provider omits thinking because these models
do not support the field.
Models that require thinking use between_tools when supported, or adaptive
thinking otherwise. For the adaptive fallback, portable reasoning: 'none'
uses explicit maxReasoningEffort, then raw output_config.effort, and defaults
to low effort when neither is supplied. Unsupported disabled settings produce warnings.
If a model rejects disabled or between_tools
thinking at xhigh or max effort, the provider lowers the effort to high
with a warning.
For Anthropic models, Bedrock's maxTokens limit includes both thinking and
final text. The provider therefore adds budgetTokens to maxOutputTokens
when setting maxTokens. When maxOutputTokens is omitted, the provider adds
budgetTokens to the default of 4096.
See AI SDK UI: Chatbot for more details on how to integrate reasoning into your chatbot.
Service Tiers
Amazon Bedrock supports selecting an inference service tier per request via the serviceTier provider option.
Supported values are:
reservedprioritydefaultflex
See the Amazon Bedrock service tiers documentation for model availability and behavior.
Extended Context Window
Claude Sonnet 4 models on Amazon Bedrock support an extended context window of up to 1 million tokens when using the context-1m-2025-08-07 beta feature.
Computer Use
Via Anthropic, Amazon Bedrock provides three provider-defined tools that can be used to interact with external systems:
- Bash Tool: Allows running bash commands.
- Text Editor Tool: Provides functionality for viewing and editing text files.
- Computer Tool: Enables control of keyboard and mouse actions on a computer.
They are available via the tools property of the provider instance.
Amazon Bedrock does not support strict mode on tool definitions. Setting
strict: true on a tool will be ignored and a warning will be emitted.
Bash Tool
The Bash Tool allows running bash commands. Here's how to create and use it:
Parameters:
command(string): The bash command to run. Required unless the tool is being restarted.restart(boolean, optional): Specifying true will restart this tool.
Text Editor Tool
The Text Editor Tool provides functionality for viewing and editing text files.
For Claude 4 models (Opus & Sonnet):
For Claude 3.5 Sonnet and earlier models:
Parameters:
command('view' | 'create' | 'str_replace' | 'insert' | 'undo_edit'): The command to run. Note:undo_editis only available in Claude 3.5 Sonnet and earlier models.path(string): Absolute path to file or directory, e.g./repo/file.pyor/repo.file_text(string, optional): Required forcreatecommand, with the content of the file to be created.insert_line(number, optional): Required forinsertcommand. The line number after which to insert the new string.new_str(string, optional): New string forstr_replacecommand.insert_text(string, optional): Required forinsertcommand, containing the text to insert.old_str(string, optional): Required forstr_replacecommand, containing the string to replace.view_range(number[], optional): Optional forviewcommand to specify line range to show.
When using the Text Editor Tool, make sure to name the key in the tools object correctly:
- Claude 4 models: Use
str_replace_based_edit_tool - Claude 3.5 Sonnet and earlier: Use
str_replace_editor
Computer Tool
The Computer Tool enables control of keyboard and mouse actions on a computer:
Parameters:
action('key' | 'type' | 'mouse_move' | 'left_click' | 'left_click_drag' | 'right_click' | 'middle_click' | 'double_click' | 'screenshot' | 'cursor_position'): The action to perform.coordinate(number[], optional): Required formouse_moveandleft_click_dragactions. Specifies the (x, y) coordinates.text(string, optional): Required fortypeandkeyactions.
These tools can be used in conjunction with the anthropic.claude-3-5-sonnet-20240620-v1:0 model to enable more complex interactions and tasks.
Model Capabilities
| Model | Image Input | Object Generation | Tool Usage | Tool Streaming |
|---|---|---|---|---|
amazon.titan-tg1-large | ||||
amazon.titan-text-express-v1 | ||||
anthropic.claude-v2 | ||||
anthropic.claude-v2:1 | ||||
anthropic.claude-instant-v1 | ||||
anthropic.claude-haiku-4-5-20251001-v1:0 | ||||
anthropic.claude-sonnet-4-5-20250929-v1:0 | ||||
anthropic.claude-sonnet-4-20250514-v1:0 | ||||
anthropic.claude-opus-4-20250514-v1:0 | ||||
anthropic.claude-opus-4-1-20250805-v1:0 | ||||
anthropic.claude-3-5-sonnet-20240620-v1:0 | ||||
anthropic.claude-3-5-sonnet-20241022-v2:0 | ||||
anthropic.claude-3-sonnet-20240229-v1:0 | ||||
anthropic.claude-3-haiku-20240307-v1:0 | ||||
anthropic.claude-3-opus-20240229-v1:0 | ||||
cohere.command-text-v14 | ||||
cohere.command-light-text-v14 | ||||
cohere.command-r-v1:0 | ||||
cohere.command-r-plus-v1:0 | ||||
meta.llama3-70b-instruct-v1:0 | ||||
meta.llama3-8b-instruct-v1:0 | ||||
meta.llama3-1-405b-instruct-v1:0 | ||||
meta.llama3-1-70b-instruct-v1:0 | ||||
meta.llama3-1-8b-instruct-v1:0 | ||||
meta.llama3-2-11b-instruct-v1:0 | ||||
meta.llama3-2-1b-instruct-v1:0 | ||||
meta.llama3-2-3b-instruct-v1:0 | ||||
meta.llama3-2-90b-instruct-v1:0 | ||||
mistral.mistral-7b-instruct-v0:2 | ||||
mistral.mixtral-8x7b-instruct-v0:1 | ||||
mistral.mistral-large-2402-v1:0 | ||||
mistral.mistral-small-2402-v1:0 | ||||
openai.gpt-oss-120b-1:0 | ||||
openai.gpt-oss-20b-1:0 | ||||
amazon.titan-text-lite-v1 | ||||
us.amazon.nova-premier-v1:0 | ||||
us.amazon.nova-pro-v1:0 | ||||
us.amazon.nova-micro-v1:0 | ||||
us.amazon.nova-lite-v1:0 | ||||
us.anthropic.claude-3-sonnet-20240229-v1:0 | ||||
us.anthropic.claude-3-opus-20240229-v1:0 | ||||
us.anthropic.claude-3-haiku-20240307-v1:0 | ||||
us.anthropic.claude-3-5-sonnet-20240620-v1:0 | ||||
us.anthropic.claude-3-5-sonnet-20241022-v2:0 | ||||
us.anthropic.claude-sonnet-4-5-20250929-v1:0 | ||||
us.anthropic.claude-sonnet-4-20250514-v1:0 | ||||
us.anthropic.claude-opus-4-20250514-v1:0 | ||||
us.anthropic.claude-opus-4-1-20250805-v1:0 | ||||
us.meta.llama3-2-11b-instruct-v1:0 | ||||
us.meta.llama3-2-3b-instruct-v1:0 | ||||
us.meta.llama3-2-90b-instruct-v1:0 | ||||
us.meta.llama3-2-1b-instruct-v1:0 | ||||
us.meta.llama3-1-8b-instruct-v1:0 | ||||
us.meta.llama3-1-70b-instruct-v1:0 | ||||
us.meta.llama3-3-70b-instruct-v1:0 | ||||
us.deepseek.r1-v1:0 | ||||
us.mistral.pixtral-large-2502-v1:0 | ||||
us.meta.llama4-scout-17b-instruct-v1:0 | ||||
us.meta.llama4-maverick-17b-instruct-v1:0 |
The table above lists popular models. Please see the Amazon Bedrock docs for a full list of available models. You can also pass any available provider model ID as a string if needed.
Embedding Models
You can create models that call the Bedrock API Bedrock API
using the .embedding() factory method.
Bedrock Titan embedding model amazon.titan-embed-text-v2:0 supports several additional settings. You can pass them as an options argument:
The following optional provider options are available for Bedrock Titan embedding models:
-
dimensions: number
The number of dimensions the output embeddings should have. The following values are accepted: 1024 (default), 512, 256.
-
normalize boolean
Flag indicating whether or not to normalize the output embeddings. Defaults to true.
Nova Embedding Models
Amazon Nova embedding models support additional provider options:
The following optional provider options are available for Nova embedding models:
-
embeddingDimension number
The number of dimensions for the output embeddings. Supported values: 256, 384, 1024 (default), 3072.
-
embeddingPurpose string
The purpose of the embedding. Accepts:
GENERIC_INDEX(default),TEXT_RETRIEVAL,IMAGE_RETRIEVAL,VIDEO_RETRIEVAL,DOCUMENT_RETRIEVAL,AUDIO_RETRIEVAL,GENERIC_RETRIEVAL,CLASSIFICATION,CLUSTERING. -
truncate string
Truncation behavior when input exceeds the model's context length. Accepts:
NONE,START,END(default).
Cohere Embedding Models
Cohere embedding models on Bedrock require an inputType and support truncation:
The following provider options are available for Cohere embedding models:
-
inputType string
Input type for Cohere embedding models. Accepts:
search_document,search_query(default),classification,clustering. -
truncate string
Truncation behavior when input exceeds the model's context length. Accepts:
NONE,START,END.
Application Inference Profiles
Application inference profile ARNs do not identify their underlying model. Pass the model family when creating an embedding model so the provider can select the correct request format and batch size:
Model Capabilities
| Model | Default Dimensions | Custom Dimensions |
|---|---|---|
amazon.titan-embed-text-v1 | 1536 | |
amazon.titan-embed-text-v2:0 | 1024 | |
amazon.nova-embed-text-v2:0 | 1024 | |
cohere.embed-english-v3 | 1024 | |
cohere.embed-multilingual-v3 | 1024 | |
cohere.embed-v4:0 | 1536 |
Reranking Models
You can create models that call the Bedrock Rerank API
using the .reranking() factory method.
You can use Amazon Bedrock reranking models to rerank documents with the rerank function:
Amazon Bedrock reranking models support additional provider options that can be passed via providerOptions.bedrock:
The following provider options are available:
-
nextToken string
Token for pagination of results.
-
additionalModelRequestFields Record<string, unknown>
Additional model-specific request fields.
Model Capabilities
| Model | Text Documents | Object Documents | Top N |
|---|---|---|---|
amazon.rerank-v1:0 | |||
cohere.rerank-v3-5:0 |
Image Models
You can create models that call the Bedrock API Bedrock API
using the .image() factory method.
For more on the Amazon Nova Canvas image model, see the Nova Canvas Overview.
The amazon.nova-canvas-v1:0 model is available in the us-east-1,
eu-west-1, and ap-northeast-1 regions.
You can then generate images with the generateImage function:
You can also pass the providerOptions object to the generateImage function to customize the generation behavior. Use the AmazonBedrockImageModelOptions type to validate Amazon Bedrock image provider options. The preferred provider option key is providerOptions.amazonBedrock; the legacy providerOptions.bedrock key remains supported for backward compatibility:
The following optional provider options are available for Amazon Nova Canvas:
-
quality string
The quality level for image generation. Accepts
'standard'or'premium'. -
negativeText string
Text describing what you don't want in the generated image.
-
cfgScale number
Controls how closely the generated image adheres to the prompt. Higher values result in images that are more closely aligned to the prompt.
-
style string
Predefined visual style for image generation. Accepts one of:
3D_ANIMATED_FAMILY_FILM·DESIGN_SKETCH·FLAT_VECTOR_ILLUSTRATION·GRAPHIC_NOVEL_ILLUSTRATION·MAXIMALISM·MIDCENTURY_RETRO·PHOTOREALISM·SOFT_DIGITAL_PAINTING.
Documentation for additional settings can be found within the Amazon Bedrock User Guide for Amazon Nova Documentation.
Image Editing
Amazon Nova Canvas supports several image editing task types. When you provide input images via prompt.images, the model automatically detects the appropriate editing mode, or you can explicitly specify the taskType in provider options.
Image Variation
Create variations of an existing image while maintaining its core characteristics:
-
similarityStrength number
Controls how similar the output is to the input image. Values range from 0.2 to 1.0, where higher values produce results closer to the original.
Inpainting
Edit specific parts of an image. You can define the area to modify using either a mask image or a text prompt:
Using a mask prompt (text-based selection):
Using a mask image:
-
maskPrompt string
A text description of the area to modify. The model will automatically identify and mask the described region.
Outpainting
Extend an image beyond its original boundaries:
-
outPaintingMode string
Controls how the outpainting is performed. Accepts
'DEFAULT'or'PRECISE'.
Background Removal
Remove the background from an image:
Background removal does not require a text prompt - only the input image is needed.
Image Editing Provider Options
The following additional provider options are available for image editing:
-
taskType string
Explicitly set the editing task type. Accepts
'TEXT_IMAGE'(default for text-only),'IMAGE_VARIATION','INPAINTING','OUTPAINTING', or'BACKGROUND_REMOVAL'. When images are provided without an explicit taskType, the model defaults to'IMAGE_VARIATION'(or'INPAINTING'if a mask is provided). -
maskPrompt string
Text description of the area to modify (for inpainting/outpainting). Alternative to providing a mask image.
-
similarityStrength number
For
IMAGE_VARIATION: Controls similarity to the original (0.2-1.0). -
outPaintingMode string
For
OUTPAINTING: Controls the outpainting behavior ('DEFAULT'or'PRECISE').
Image Model Settings
You can customize the generation behavior with optional options:
-
maxImagesPerCall number
Override the maximum number of images generated per API call. Default can vary by model, with 5 as a common default.
Model Capabilities
The Amazon Nova Canvas model supports custom sizes with constraints as follows:
- Each side must be between 320-4096 pixels, inclusive.
- Each side must be evenly divisible by 16.
- The aspect ratio must be between 1:4 and 4:1. That is, one side can't be more than 4 times longer than the other side.
- The total pixel count must be less than 4,194,304.
For more, see Image generation access and usage.
| Model | Sizes | Image Editing | Mask Input | Max Images Per Call |
|---|---|---|---|---|
amazon.nova-canvas-v1:0 | 320–4096 pixels per side, divisible by 16; aspect ratio 1:4 to 4:1; fewer than 4,194,304 pixels. | 5 |
Response Headers
The Amazon Bedrock provider will return the response headers associated with network requests made of the Bedrock servers.
Below is sample output where you can see the x-amzn-requestid header. This can
be useful for correlating Bedrock API calls with requests made by the AI SDK:
This information is also available with streamText:
With sample output as:
Bedrock Anthropic Provider Usage
The Bedrock Anthropic provider offers support for Anthropic's Claude models through Amazon Bedrock's native InvokeModel API. This provides full feature parity with the Anthropic API, including features that may not be available through the Converse API (such as stop_sequence in streaming responses).
For more information on Claude models available on Amazon Bedrock, see Claude on Amazon Bedrock.
Provider Instance
You can import the default provider instance bedrockAnthropic from @ai-sdk/amazon-bedrock/anthropic:
If you need a customized setup, you can import createBedrockAnthropic from @ai-sdk/amazon-bedrock/anthropic and create a provider instance with your settings:
Provider Settings
You can use the following optional settings to customize the Bedrock Anthropic provider instance:
-
region string
The AWS region that you want to use for the API calls. It uses the
AWS_REGIONenvironment variable by default. -
accessKeyId string
The AWS access key ID that you want to use for the API calls. It uses the
AWS_ACCESS_KEY_IDenvironment variable by default. -
secretAccessKey string
The AWS secret access key that you want to use for the API calls. It uses the
AWS_SECRET_ACCESS_KEYenvironment variable by default. -
sessionToken string
Optional. For temporary credentials. With both access keys set as strings, pass the token here if needed;
AWS_SESSION_TOKENfrom the environment is not used. If either key is taken from the environment, omittingsessionTokenallowsAWS_SESSION_TOKEN. -
apiKey string | PromiseLike<string> | (() => string | PromiseLike<string>)
Pass a callback to resolve the API key before each request. The callback can return a string or a promise; token caching and renewal are handled by the caller. A callback or promise selects API key authentication and must resolve to a non-empty string; lookup failures do not switch authentication methods.
API key for authenticating requests using Bearer token authentication. When provided, this will be used instead of AWS SigV4 authentication. It uses the
AWS_BEARER_TOKEN_BEDROCKenvironment variable by default. -
baseURL string
Base URL for the Bedrock API calls. Useful for custom endpoints or proxy configurations. When omitted, the provider also supports the service-specific
AWS_ENDPOINT_URL_BEDROCK_RUNTIMEenvironment variable, which takes precedence over the globalAWS_ENDPOINT_URLenvironment variable. -
headers Resolvable<Record<string, string | undefined>>
Headers to include in the requests.
-
fetch (input: RequestInfo, init?: RequestInit) => Promise<Response>
Custom fetch implementation. You can use it as a middleware to intercept requests, or to provide a custom fetch implementation for e.g. testing.
-
credentialProvider () => PromiseLike<BedrockCredentials>
The AWS credential provider to use for the Bedrock provider to get dynamic credentials similar to the AWS SDK. Setting a provider here will cause its credential values to be used instead of the
accessKeyId,secretAccessKey, andsessionTokensettings.
Language Models
You can create models that call the Anthropic Messages API using the provider instance.
The first argument is the model id, e.g. us.anthropic.claude-sonnet-5-5.
You can use Bedrock Anthropic language models to generate text with the generateText function:
Provider Options
The following optional provider options are available for Bedrock Anthropic models:
-
metadataobjectOptional. Metadata to include with the request. See the Anthropic API documentation for details.
userIdstring - An external identifier for the end-user.
Cache Control
In the messages and message parts, you can use the providerOptions property to set cache control breakpoints.
You need to set the anthropic property in the providerOptions object to { cacheControl: { type: 'ephemeral' } } to set a cache control breakpoint.
Cache control requires a minimum of 1024 tokens before the cache checkpoint. See the Amazon Bedrock prompt caching documentation for details on supported models and limits.
Computer Use
The Bedrock Anthropic provider supports Anthropic's computer use tools:
- Bash Tool: Allows running bash commands.
- Text Editor Tool: Provides functionality for viewing and editing text files.
- Computer Tool: Enables control of keyboard and mouse actions on a computer.
They are available via the tools property of the provider instance.
Computer use tools require Claude 3.7 Sonnet or newer models. Claude 3.5 Sonnet v2 does not support these tools.
Bash Tool
Text Editor Tool
Computer Tool
Reasoning
Anthropic has reasoning support for Claude 3.7 and Claude 4 models on Bedrock, including:
us.anthropic.claude-fable-5-2us.anthropic.claude-fable-5-1us.anthropic.claude-sonnet-5-5us.anthropic.claude-sonnet-5us.anthropic.claude-fable-5us.anthropic.claude-opus-4-8us.anthropic.claude-opus-4-7us.anthropic.claude-opus-4-6-v1us.anthropic.claude-opus-4-5-20251101-v1:0us.anthropic.claude-sonnet-4-5-20250929-v1:0us.anthropic.claude-opus-4-20250514-v1:0us.anthropic.claude-sonnet-4-20250514-v1:0us.anthropic.claude-opus-4-1-20250805-v1:0us.anthropic.claude-haiku-4-5-20251001-v1:0
You can enable it using the thinking provider option and specifying a thinking budget in tokens.
Anthropic's max_tokens limit includes both thinking and final text. To reserve
room for both, the provider adds budgetTokens to maxOutputTokens when
setting max_tokens. For known models, the combined value is capped at the
model's maximum output token limit.
See AI SDK UI: Chatbot for more details on how to integrate reasoning into your chatbot.
Model Capabilities
| Model | Image Input | Object Generation | Tool Usage | Computer Use | Reasoning |
|---|---|---|---|---|---|
us.anthropic.claude-fable-5-1 | |||||
us.anthropic.claude-fable-5-2 | |||||
us.anthropic.claude-sonnet-5-5 | |||||
us.anthropic.claude-sonnet-5 | |||||
us.anthropic.claude-haiku-5-5 | |||||
us.anthropic.claude-fable-5 | |||||
us.anthropic.claude-opus-4-8 | |||||
us.anthropic.claude-opus-4-7 | |||||
us.anthropic.claude-opus-4-6-v1 | |||||
us.anthropic.claude-opus-4-5-20251101-v1:0 | |||||
us.anthropic.claude-sonnet-4-5-20250929-v1:0 | |||||
us.anthropic.claude-opus-4-20250514-v1:0 | |||||
us.anthropic.claude-sonnet-4-20250514-v1:0 | |||||
us.anthropic.claude-opus-4-1-20250805-v1:0 | |||||
us.anthropic.claude-haiku-4-5-20251001-v1:0 | |||||
us.anthropic.claude-3-5-sonnet-20241022-v2:0 |
The Bedrock Anthropic provider uses the native InvokeModel API and supports all features available in the Anthropic API, except for the Files API and MCP Connector which are not supported on Bedrock.
Bedrock Mantle Provider Usage
The Bedrock Mantle provider offers access to OpenAI-compatible models through Amazon Bedrock's Mantle API. It automatically selects the /v1 or /openai/v1 route based on the model. These endpoints support the Chat Completions API and the Responses API, and provide access to models that may only be available through the Mantle endpoint (such as openai.gpt-oss-120b).
Provider Instance
You can import the default provider instance bedrockMantle from @ai-sdk/amazon-bedrock/mantle:
If you need a customized setup, you can import createBedrockMantle from @ai-sdk/amazon-bedrock/mantle and create a provider instance with your settings:
Provider Settings
You can use the following optional settings to customize the Bedrock Mantle provider instance:
-
region string
The AWS region that you want to use for the API calls. It uses the
AWS_REGIONenvironment variable by default. -
accessKeyId string
The AWS access key ID that you want to use for the API calls. It uses the
AWS_ACCESS_KEY_IDenvironment variable by default. -
secretAccessKey string
The AWS secret access key that you want to use for the API calls. It uses the
AWS_SECRET_ACCESS_KEYenvironment variable by default. -
sessionToken string
Optional. The AWS session token that you want to use for the API calls. It uses the
AWS_SESSION_TOKENenvironment variable by default. -
apiKey string | PromiseLike<string> | (() => string | PromiseLike<string>)
Pass a callback to resolve the API key before each request. The callback can return a string or a promise; token caching and renewal are handled by the caller. A callback or promise selects API key authentication and must resolve to a non-empty string; lookup failures do not switch authentication methods.
API key for authenticating requests using Bearer token authentication. When provided, this will be used instead of AWS SigV4 authentication. It uses the
AWS_BEARER_TOKEN_BEDROCKenvironment variable by default. -
baseURL string
Base URL for the Bedrock Mantle API calls. By default, the provider selects either
https://bedrock-mantle.{region}.api.aws/v1orhttps://bedrock-mantle.{region}.api.aws/openai/v1based on the model. -
headers Record<string, string | undefined>
Headers to include in the requests.
-
fetch (input: RequestInfo, init?: RequestInit) => Promise<Response>
Custom fetch implementation. You can use it as a middleware to intercept requests, or to provide a custom fetch implementation for e.g. testing.
-
credentialProvider () => PromiseLike<BedrockCredentials>
The AWS credential provider to use for the Bedrock provider to get dynamic credentials similar to the AWS SDK. Setting a provider here will cause its credential values to be used instead of the
accessKeyId,secretAccessKey, andsessionTokensettings.
Language Models
You can create models using the Mantle provider instance. By default, the provider creates models using the Chat Completions API, which has the broadest model support on Mantle:
You can also explicitly select the Chat Completions API or the Responses API:
Note that not all Mantle models support the Responses API. See the model capabilities table below.
You can use Bedrock Mantle language models to generate text with the generateText function:
Or stream text with the streamText function:
Provider Options
Since the Mantle API is OpenAI-compatible, provider options are passed under the openai namespace:
Model Capabilities
| Model | Chat Completions | Responses API |
|---|---|---|
openai.gpt-oss-20b | ||
openai.gpt-oss-120b | ||
openai.gpt-oss-safeguard-20b | ||
openai.gpt-oss-safeguard-120b |
The Bedrock Mantle provider uses the OpenAI-compatible API format. Models available through the Mantle endpoint may differ from those available through the Bedrock Converse API. Use the Mantle endpoint's model listing API to discover all available models.
Migrating to @ai-sdk/amazon-bedrock 2.x
The Amazon Bedrock provider was rewritten in version 2.x to remove the
dependency on the @aws-sdk/client-bedrock-runtime package.
The bedrockOptions provider setting previously available has been removed. If
you were using the bedrockOptions object, you should now use the region,
accessKeyId, secretAccessKey, and sessionToken settings directly instead.
Static IAM user keys do not require sessionToken.
Provider Option Reference
The following option definitions are generated from the provider catalog.
| Operation | Option | Description |
|---|---|---|
| language | structuredOutputMode | Select native structured output or JSON tools. |
| language | additionalModelRequestFields | Additional model-specific inference fields. |
| language | reasoningConfig | Thinking mode, budget, effort, and display. |
| language | anthropicBeta | Anthropic beta features to enable. |
| language | serviceTier | Request service tier. |
| language | requestMetadata | Invocation logging and cost attribution metadata. |
| embedding | dimensions | Titan v2 output dimensions. |
| embedding | normalize | Normalize Titan v2 embeddings. |
| embedding | embeddingDimension | Nova output dimensions. |
| embedding | embeddingPurpose | Nova embedding purpose. |
| embedding | inputType | Cohere embedding input type. |
| embedding | truncate | Cohere or Nova truncation mode. |
| embedding | outputDimension | Cohere v4 output dimensions. |
| image | quality | Image quality. |
| image | negativeText | Content to avoid. |
| image | cfgScale | Prompt adherence scale. |
| image | style | Image generation style. |
| image | taskType | Generation, variation, inpainting, outpainting, or background removal. |
| image | maskPrompt | Text describing the region to mask. |
| image | outPaintingMode | Outpainting mode. |
| image | similarityStrength | Similarity to reference images. |
| reranking | nextToken | Pagination token. |
| reranking | additionalModelRequestFields | Additional model-specific reranking fields. |