Amazon Bedrock Runtime 2023-09-30
- Client: Aws\BedrockRuntime\BedrockRuntimeClient
- Service ID: bedrock-runtime
- Version: 2023-09-30
This page describes the parameters and results for the operations of the Amazon Bedrock Runtime (2023-09-30), and shows how to use the Aws\BedrockRuntime\BedrockRuntimeClient object to call the described operations. This documentation is specific to the 2023-09-30 API version of the service.
Operation Summary
Each of the following operations can be created from a client using
$client->getCommand('CommandName')
, where "CommandName" is the
name of one of the following operations. Note: a command is a value that
encapsulates an operation and the parameters used to create an HTTP request.
You can also create and send a command immediately using the magic methods
available on a client object: $client->commandName(/* parameters */)
.
You can send the command asynchronously (returning a promise) by appending the
word "Async" to the operation name: $client->commandNameAsync(/* parameters */)
.
- ApplyGuardrail ( array $params = [] )
- The action to apply a guardrail.
- Converse ( array $params = [] )
- Sends messages to the specified Amazon Bedrock model.
- ConverseStream ( array $params = [] )
- Sends messages to the specified Amazon Bedrock model and returns the response in a stream.
- InvokeModel ( array $params = [] )
- Invokes the specified Amazon Bedrock model to run inference using the prompt and inference parameters provided in the request body.
- InvokeModelWithResponseStream ( array $params = [] )
- Invoke the specified Amazon Bedrock model to run inference using the prompt and inference parameters provided in the request body.
Operations
ApplyGuardrail
$result = $client->applyGuardrail
([/* ... */]); $promise = $client->applyGuardrailAsync
([/* ... */]);
The action to apply a guardrail.
Parameter Syntax
$result = $client->applyGuardrail([ 'content' => [ // REQUIRED [ 'text' => [ 'qualifiers' => ['<string>', ...], 'text' => '<string>', // REQUIRED ], ], // ... ], 'guardrailIdentifier' => '<string>', // REQUIRED 'guardrailVersion' => '<string>', // REQUIRED 'source' => 'INPUT|OUTPUT', // REQUIRED ]);
Parameter Details
Members
- content
-
- Required: Yes
- Type: Array of GuardrailContentBlock structures
The content details used in the request to apply the guardrail.
- guardrailIdentifier
-
- Required: Yes
- Type: string
The guardrail identifier used in the request to apply the guardrail.
- guardrailVersion
-
- Required: Yes
- Type: string
The guardrail version used in the request to apply the guardrail.
- source
-
- Required: Yes
- Type: string
The source of data used in the request to apply the guardrail.
Result Syntax
[ 'action' => 'NONE|GUARDRAIL_INTERVENED', 'assessments' => [ [ 'contentPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED', 'confidence' => 'NONE|LOW|MEDIUM|HIGH', 'filterStrength' => 'NONE|LOW|MEDIUM|HIGH', 'type' => 'INSULTS|HATE|SEXUAL|VIOLENCE|MISCONDUCT|PROMPT_ATTACK', ], // ... ], ], 'contextualGroundingPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED|NONE', 'score' => <float>, 'threshold' => <float>, 'type' => 'GROUNDING|RELEVANCE', ], // ... ], ], 'invocationMetrics' => [ 'guardrailCoverage' => [ 'textCharacters' => [ 'guarded' => <integer>, 'total' => <integer>, ], ], 'guardrailProcessingLatency' => <integer>, 'usage' => [ 'contentPolicyUnits' => <integer>, 'contextualGroundingPolicyUnits' => <integer>, 'sensitiveInformationPolicyFreeUnits' => <integer>, 'sensitiveInformationPolicyUnits' => <integer>, 'topicPolicyUnits' => <integer>, 'wordPolicyUnits' => <integer>, ], ], 'sensitiveInformationPolicy' => [ 'piiEntities' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'type' => 'ADDRESS|AGE|AWS_ACCESS_KEY|AWS_SECRET_KEY|CA_HEALTH_NUMBER|CA_SOCIAL_INSURANCE_NUMBER|CREDIT_DEBIT_CARD_CVV|CREDIT_DEBIT_CARD_EXPIRY|CREDIT_DEBIT_CARD_NUMBER|DRIVER_ID|EMAIL|INTERNATIONAL_BANK_ACCOUNT_NUMBER|IP_ADDRESS|LICENSE_PLATE|MAC_ADDRESS|NAME|PASSWORD|PHONE|PIN|SWIFT_CODE|UK_NATIONAL_HEALTH_SERVICE_NUMBER|UK_NATIONAL_INSURANCE_NUMBER|UK_UNIQUE_TAXPAYER_REFERENCE_NUMBER|URL|USERNAME|US_BANK_ACCOUNT_NUMBER|US_BANK_ROUTING_NUMBER|US_INDIVIDUAL_TAX_IDENTIFICATION_NUMBER|US_PASSPORT_NUMBER|US_SOCIAL_SECURITY_NUMBER|VEHICLE_IDENTIFICATION_NUMBER', ], // ... ], 'regexes' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'name' => '<string>', 'regex' => '<string>', ], // ... ], ], 'topicPolicy' => [ 'topics' => [ [ 'action' => 'BLOCKED', 'name' => '<string>', 'type' => 'DENY', ], // ... ], ], 'wordPolicy' => [ 'customWords' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', ], // ... ], 'managedWordLists' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', 'type' => 'PROFANITY', ], // ... ], ], ], // ... ], 'guardrailCoverage' => [ 'textCharacters' => [ 'guarded' => <integer>, 'total' => <integer>, ], ], 'outputs' => [ [ 'text' => '<string>', ], // ... ], 'usage' => [ 'contentPolicyUnits' => <integer>, 'contextualGroundingPolicyUnits' => <integer>, 'sensitiveInformationPolicyFreeUnits' => <integer>, 'sensitiveInformationPolicyUnits' => <integer>, 'topicPolicyUnits' => <integer>, 'wordPolicyUnits' => <integer>, ], ]
Result Details
Members
- action
-
- Required: Yes
- Type: string
The action taken in the response from the guardrail.
- assessments
-
- Required: Yes
- Type: Array of GuardrailAssessment structures
The assessment details in the response from the guardrail.
- guardrailCoverage
-
- Type: GuardrailCoverage structure
The guardrail coverage details in the apply guardrail response.
- outputs
-
- Required: Yes
- Type: Array of GuardrailOutputContent structures
The output details in the response from the guardrail.
- usage
-
- Required: Yes
- Type: GuardrailUsage structure
The usage details in the response from the guardrail.
Errors
- AccessDeniedException:
The request is denied because of missing access permissions.
- ResourceNotFoundException:
The specified resource ARN was not found. Check the ARN and try your request again.
- ThrottlingException:
Your request was throttled because of service-wide limitations. Resubmit your request later or in a different region. You can also purchase Provisioned Throughput to increase the rate or number of tokens you can process.
- InternalServerException:
An internal server error occurred. Retry your request.
- ValidationException:
Input validation failed. Check your request parameters and retry the request.
- ServiceQuotaExceededException:
Your request exceeds the service quota for your account. You can view your quotas at Viewing service quotas. You can resubmit your request later.
Converse
$result = $client->converse
([/* ... */]); $promise = $client->converseAsync
([/* ... */]);
Sends messages to the specified Amazon Bedrock model. Converse
provides a consistent interface that works with all models that support messages. This allows you to write code once and use it with different models. If a model has unique inference parameters, you can also pass those unique parameters to the model.
Amazon Bedrock doesn't store any text, images, or documents that you provide as content. The data is only used to generate the response.
For information about the Converse API, see Use the Converse API in the Amazon Bedrock User Guide. To use a guardrail, see Use a guardrail with the Converse API in the Amazon Bedrock User Guide. To use a tool with a model, see Tool use (Function calling) in the Amazon Bedrock User Guide
For example code, see Converse API examples in the Amazon Bedrock User Guide.
This operation requires permission for the bedrock:InvokeModel
action.
Parameter Syntax
$result = $client->converse([ 'additionalModelRequestFields' => [ ], 'additionalModelResponseFieldPaths' => ['<string>', ...], 'guardrailConfig' => [ 'guardrailIdentifier' => '<string>', // REQUIRED 'guardrailVersion' => '<string>', // REQUIRED 'trace' => 'enabled|disabled', ], 'inferenceConfig' => [ 'maxTokens' => <integer>, 'stopSequences' => ['<string>', ...], 'temperature' => <float>, 'topP' => <float>, ], 'messages' => [ // REQUIRED [ 'content' => [ // REQUIRED [ 'document' => [ 'format' => 'pdf|csv|doc|docx|xls|xlsx|html|txt|md', // REQUIRED 'name' => '<string>', // REQUIRED 'source' => [ // REQUIRED 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'guardContent' => [ 'text' => [ 'qualifiers' => ['<string>', ...], 'text' => '<string>', // REQUIRED ], ], 'image' => [ 'format' => 'png|jpeg|gif|webp', // REQUIRED 'source' => [ // REQUIRED 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'text' => '<string>', 'toolResult' => [ 'content' => [ // REQUIRED [ 'document' => [ 'format' => 'pdf|csv|doc|docx|xls|xlsx|html|txt|md', // REQUIRED 'name' => '<string>', // REQUIRED 'source' => [ // REQUIRED 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'image' => [ 'format' => 'png|jpeg|gif|webp', // REQUIRED 'source' => [ // REQUIRED 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'json' => [ ], 'text' => '<string>', ], // ... ], 'status' => 'success|error', 'toolUseId' => '<string>', // REQUIRED ], 'toolUse' => [ 'input' => [ // REQUIRED ], 'name' => '<string>', // REQUIRED 'toolUseId' => '<string>', // REQUIRED ], ], // ... ], 'role' => 'user|assistant', // REQUIRED ], // ... ], 'modelId' => '<string>', // REQUIRED 'system' => [ [ 'guardContent' => [ 'text' => [ 'qualifiers' => ['<string>', ...], 'text' => '<string>', // REQUIRED ], ], 'text' => '<string>', ], // ... ], 'toolConfig' => [ 'toolChoice' => [ 'any' => [ ], 'auto' => [ ], 'tool' => [ 'name' => '<string>', // REQUIRED ], ], 'tools' => [ // REQUIRED [ 'toolSpec' => [ 'description' => '<string>', 'inputSchema' => [ // REQUIRED 'json' => [ ], ], 'name' => '<string>', // REQUIRED ], ], // ... ], ], ]);
Parameter Details
Members
- additionalModelRequestFields
-
- Type: document (null|bool|string|numeric) or an (array|associative array) whose members are all valid documents
Additional inference parameters that the model supports, beyond the base set of inference parameters that
Converse
supports in theinferenceConfig
field. For more information, see Model parameters. - additionalModelResponseFieldPaths
-
- Type: Array of strings
Additional model parameters field paths to return in the response.
Converse
returns the requested fields as a JSON Pointer object in theadditionalModelResponseFields
field. The following is example JSON foradditionalModelResponseFieldPaths
.[ "/stop_sequence" ]
For information about the JSON Pointer syntax, see the Internet Engineering Task Force (IETF) documentation.
Converse
rejects an empty JSON Pointer or incorrectly structured JSON Pointer with a400
error code. if the JSON Pointer is valid, but the requested field is not in the model response, it is ignored byConverse
. - guardrailConfig
-
- Type: GuardrailConfiguration structure
Configuration information for a guardrail that you want to use in the request.
- inferenceConfig
-
- Type: InferenceConfiguration structure
Inference parameters to pass to the model.
Converse
supports a base set of inference parameters. If you need to pass additional parameters that the model supports, use theadditionalModelRequestFields
request field. - messages
-
- Required: Yes
- Type: Array of Message structures
The messages that you want to send to the model.
- modelId
-
- Required: Yes
- Type: string
The identifier for the model that you want to call.
The
modelId
to provide depends on the type of model or throughput that you use:-
If you use a base model, specify the model ID or its ARN. For a list of model IDs for base models, see Amazon Bedrock base model IDs (on-demand throughput) in the Amazon Bedrock User Guide.
-
If you use an inference profile, specify the inference profile ID or its ARN. For a list of inference profile IDs, see Supported Regions and models for cross-region inference in the Amazon Bedrock User Guide.
-
If you use a provisioned model, specify the ARN of the Provisioned Throughput. For more information, see Run inference using a Provisioned Throughput in the Amazon Bedrock User Guide.
-
If you use a custom model, first purchase Provisioned Throughput for it. Then specify the ARN of the resulting provisioned model. For more information, see Use a custom model in Amazon Bedrock in the Amazon Bedrock User Guide.
The Converse API doesn't support imported models.
- system
-
- Type: Array of SystemContentBlock structures
A system prompt to pass to the model.
- toolConfig
-
- Type: ToolConfiguration structure
Configuration information for the tools that the model can use when generating a response.
This field is only supported by Anthropic Claude 3, Cohere Command R, Cohere Command R+, and Mistral Large models.
Result Syntax
[ 'additionalModelResponseFields' => [ ], 'metrics' => [ 'latencyMs' => <integer>, ], 'output' => [ 'message' => [ 'content' => [ [ 'document' => [ 'format' => 'pdf|csv|doc|docx|xls|xlsx|html|txt|md', 'name' => '<string>', 'source' => [ 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'guardContent' => [ 'text' => [ 'qualifiers' => ['<string>', ...], 'text' => '<string>', ], ], 'image' => [ 'format' => 'png|jpeg|gif|webp', 'source' => [ 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'text' => '<string>', 'toolResult' => [ 'content' => [ [ 'document' => [ 'format' => 'pdf|csv|doc|docx|xls|xlsx|html|txt|md', 'name' => '<string>', 'source' => [ 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'image' => [ 'format' => 'png|jpeg|gif|webp', 'source' => [ 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'json' => [ ], 'text' => '<string>', ], // ... ], 'status' => 'success|error', 'toolUseId' => '<string>', ], 'toolUse' => [ 'input' => [ ], 'name' => '<string>', 'toolUseId' => '<string>', ], ], // ... ], 'role' => 'user|assistant', ], ], 'stopReason' => 'end_turn|tool_use|max_tokens|stop_sequence|guardrail_intervened|content_filtered', 'trace' => [ 'guardrail' => [ 'inputAssessment' => [ '<String>' => [ 'contentPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED', 'confidence' => 'NONE|LOW|MEDIUM|HIGH', 'filterStrength' => 'NONE|LOW|MEDIUM|HIGH', 'type' => 'INSULTS|HATE|SEXUAL|VIOLENCE|MISCONDUCT|PROMPT_ATTACK', ], // ... ], ], 'contextualGroundingPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED|NONE', 'score' => <float>, 'threshold' => <float>, 'type' => 'GROUNDING|RELEVANCE', ], // ... ], ], 'invocationMetrics' => [ 'guardrailCoverage' => [ 'textCharacters' => [ 'guarded' => <integer>, 'total' => <integer>, ], ], 'guardrailProcessingLatency' => <integer>, 'usage' => [ 'contentPolicyUnits' => <integer>, 'contextualGroundingPolicyUnits' => <integer>, 'sensitiveInformationPolicyFreeUnits' => <integer>, 'sensitiveInformationPolicyUnits' => <integer>, 'topicPolicyUnits' => <integer>, 'wordPolicyUnits' => <integer>, ], ], 'sensitiveInformationPolicy' => [ 'piiEntities' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'type' => 'ADDRESS|AGE|AWS_ACCESS_KEY|AWS_SECRET_KEY|CA_HEALTH_NUMBER|CA_SOCIAL_INSURANCE_NUMBER|CREDIT_DEBIT_CARD_CVV|CREDIT_DEBIT_CARD_EXPIRY|CREDIT_DEBIT_CARD_NUMBER|DRIVER_ID|EMAIL|INTERNATIONAL_BANK_ACCOUNT_NUMBER|IP_ADDRESS|LICENSE_PLATE|MAC_ADDRESS|NAME|PASSWORD|PHONE|PIN|SWIFT_CODE|UK_NATIONAL_HEALTH_SERVICE_NUMBER|UK_NATIONAL_INSURANCE_NUMBER|UK_UNIQUE_TAXPAYER_REFERENCE_NUMBER|URL|USERNAME|US_BANK_ACCOUNT_NUMBER|US_BANK_ROUTING_NUMBER|US_INDIVIDUAL_TAX_IDENTIFICATION_NUMBER|US_PASSPORT_NUMBER|US_SOCIAL_SECURITY_NUMBER|VEHICLE_IDENTIFICATION_NUMBER', ], // ... ], 'regexes' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'name' => '<string>', 'regex' => '<string>', ], // ... ], ], 'topicPolicy' => [ 'topics' => [ [ 'action' => 'BLOCKED', 'name' => '<string>', 'type' => 'DENY', ], // ... ], ], 'wordPolicy' => [ 'customWords' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', ], // ... ], 'managedWordLists' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', 'type' => 'PROFANITY', ], // ... ], ], ], // ... ], 'modelOutput' => ['<string>', ...], 'outputAssessments' => [ '<String>' => [ [ 'contentPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED', 'confidence' => 'NONE|LOW|MEDIUM|HIGH', 'filterStrength' => 'NONE|LOW|MEDIUM|HIGH', 'type' => 'INSULTS|HATE|SEXUAL|VIOLENCE|MISCONDUCT|PROMPT_ATTACK', ], // ... ], ], 'contextualGroundingPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED|NONE', 'score' => <float>, 'threshold' => <float>, 'type' => 'GROUNDING|RELEVANCE', ], // ... ], ], 'invocationMetrics' => [ 'guardrailCoverage' => [ 'textCharacters' => [ 'guarded' => <integer>, 'total' => <integer>, ], ], 'guardrailProcessingLatency' => <integer>, 'usage' => [ 'contentPolicyUnits' => <integer>, 'contextualGroundingPolicyUnits' => <integer>, 'sensitiveInformationPolicyFreeUnits' => <integer>, 'sensitiveInformationPolicyUnits' => <integer>, 'topicPolicyUnits' => <integer>, 'wordPolicyUnits' => <integer>, ], ], 'sensitiveInformationPolicy' => [ 'piiEntities' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'type' => 'ADDRESS|AGE|AWS_ACCESS_KEY|AWS_SECRET_KEY|CA_HEALTH_NUMBER|CA_SOCIAL_INSURANCE_NUMBER|CREDIT_DEBIT_CARD_CVV|CREDIT_DEBIT_CARD_EXPIRY|CREDIT_DEBIT_CARD_NUMBER|DRIVER_ID|EMAIL|INTERNATIONAL_BANK_ACCOUNT_NUMBER|IP_ADDRESS|LICENSE_PLATE|MAC_ADDRESS|NAME|PASSWORD|PHONE|PIN|SWIFT_CODE|UK_NATIONAL_HEALTH_SERVICE_NUMBER|UK_NATIONAL_INSURANCE_NUMBER|UK_UNIQUE_TAXPAYER_REFERENCE_NUMBER|URL|USERNAME|US_BANK_ACCOUNT_NUMBER|US_BANK_ROUTING_NUMBER|US_INDIVIDUAL_TAX_IDENTIFICATION_NUMBER|US_PASSPORT_NUMBER|US_SOCIAL_SECURITY_NUMBER|VEHICLE_IDENTIFICATION_NUMBER', ], // ... ], 'regexes' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'name' => '<string>', 'regex' => '<string>', ], // ... ], ], 'topicPolicy' => [ 'topics' => [ [ 'action' => 'BLOCKED', 'name' => '<string>', 'type' => 'DENY', ], // ... ], ], 'wordPolicy' => [ 'customWords' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', ], // ... ], 'managedWordLists' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', 'type' => 'PROFANITY', ], // ... ], ], ], // ... ], // ... ], ], ], 'usage' => [ 'inputTokens' => <integer>, 'outputTokens' => <integer>, 'totalTokens' => <integer>, ], ]
Result Details
Members
- additionalModelResponseFields
-
- Type: document (null|bool|string|numeric) or an (array|associative array) whose members are all valid documents
Additional fields in the response that are unique to the model.
- metrics
-
- Required: Yes
- Type: ConverseMetrics structure
Metrics for the call to
Converse
. - output
-
- Required: Yes
- Type: ConverseOutput structure
The result from the call to
Converse
. - stopReason
-
- Required: Yes
- Type: string
The reason why the model stopped generating output.
- trace
-
- Type: ConverseTrace structure
A trace object that contains information about the Guardrail behavior.
- usage
-
- Required: Yes
- Type: TokenUsage structure
The total number of tokens used in the call to
Converse
. The total includes the tokens input to the model and the tokens generated by the model.
Errors
- AccessDeniedException:
The request is denied because of missing access permissions.
- ResourceNotFoundException:
The specified resource ARN was not found. Check the ARN and try your request again.
- ThrottlingException:
Your request was throttled because of service-wide limitations. Resubmit your request later or in a different region. You can also purchase Provisioned Throughput to increase the rate or number of tokens you can process.
- ModelTimeoutException:
The request took too long to process. Processing time exceeded the model timeout length.
- InternalServerException:
An internal server error occurred. Retry your request.
- ServiceUnavailableException:
The service isn't currently available. Try again later.
- ValidationException:
Input validation failed. Check your request parameters and retry the request.
- ModelNotReadyException:
The model specified in the request is not ready to serve inference requests. The AWS SDK will automatically retry the operation up to 5 times. For information about configuring automatic retries, see Retry behavior in the AWS SDKs and Tools reference guide.
- ModelErrorException:
The request failed due to an error while processing the model.
ConverseStream
$result = $client->converseStream
([/* ... */]); $promise = $client->converseStreamAsync
([/* ... */]);
Sends messages to the specified Amazon Bedrock model and returns the response in a stream. ConverseStream
provides a consistent API that works with all Amazon Bedrock models that support messages. This allows you to write code once and use it with different models. Should a model have unique inference parameters, you can also pass those unique parameters to the model.
To find out if a model supports streaming, call GetFoundationModel and check the responseStreamingSupported
field in the response.
The CLI doesn't support streaming operations in Amazon Bedrock, including ConverseStream
.
Amazon Bedrock doesn't store any text, images, or documents that you provide as content. The data is only used to generate the response.
For information about the Converse API, see Use the Converse API in the Amazon Bedrock User Guide. To use a guardrail, see Use a guardrail with the Converse API in the Amazon Bedrock User Guide. To use a tool with a model, see Tool use (Function calling) in the Amazon Bedrock User Guide
For example code, see Conversation streaming example in the Amazon Bedrock User Guide.
This operation requires permission for the bedrock:InvokeModelWithResponseStream
action.
Parameter Syntax
$result = $client->converseStream([ 'additionalModelRequestFields' => [ ], 'additionalModelResponseFieldPaths' => ['<string>', ...], 'guardrailConfig' => [ 'guardrailIdentifier' => '<string>', // REQUIRED 'guardrailVersion' => '<string>', // REQUIRED 'streamProcessingMode' => 'sync|async', 'trace' => 'enabled|disabled', ], 'inferenceConfig' => [ 'maxTokens' => <integer>, 'stopSequences' => ['<string>', ...], 'temperature' => <float>, 'topP' => <float>, ], 'messages' => [ // REQUIRED [ 'content' => [ // REQUIRED [ 'document' => [ 'format' => 'pdf|csv|doc|docx|xls|xlsx|html|txt|md', // REQUIRED 'name' => '<string>', // REQUIRED 'source' => [ // REQUIRED 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'guardContent' => [ 'text' => [ 'qualifiers' => ['<string>', ...], 'text' => '<string>', // REQUIRED ], ], 'image' => [ 'format' => 'png|jpeg|gif|webp', // REQUIRED 'source' => [ // REQUIRED 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'text' => '<string>', 'toolResult' => [ 'content' => [ // REQUIRED [ 'document' => [ 'format' => 'pdf|csv|doc|docx|xls|xlsx|html|txt|md', // REQUIRED 'name' => '<string>', // REQUIRED 'source' => [ // REQUIRED 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'image' => [ 'format' => 'png|jpeg|gif|webp', // REQUIRED 'source' => [ // REQUIRED 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], ], 'json' => [ ], 'text' => '<string>', ], // ... ], 'status' => 'success|error', 'toolUseId' => '<string>', // REQUIRED ], 'toolUse' => [ 'input' => [ // REQUIRED ], 'name' => '<string>', // REQUIRED 'toolUseId' => '<string>', // REQUIRED ], ], // ... ], 'role' => 'user|assistant', // REQUIRED ], // ... ], 'modelId' => '<string>', // REQUIRED 'system' => [ [ 'guardContent' => [ 'text' => [ 'qualifiers' => ['<string>', ...], 'text' => '<string>', // REQUIRED ], ], 'text' => '<string>', ], // ... ], 'toolConfig' => [ 'toolChoice' => [ 'any' => [ ], 'auto' => [ ], 'tool' => [ 'name' => '<string>', // REQUIRED ], ], 'tools' => [ // REQUIRED [ 'toolSpec' => [ 'description' => '<string>', 'inputSchema' => [ // REQUIRED 'json' => [ ], ], 'name' => '<string>', // REQUIRED ], ], // ... ], ], ]);
Parameter Details
Members
- additionalModelRequestFields
-
- Type: document (null|bool|string|numeric) or an (array|associative array) whose members are all valid documents
Additional inference parameters that the model supports, beyond the base set of inference parameters that
ConverseStream
supports in theinferenceConfig
field. - additionalModelResponseFieldPaths
-
- Type: Array of strings
Additional model parameters field paths to return in the response.
ConverseStream
returns the requested fields as a JSON Pointer object in theadditionalModelResponseFields
field. The following is example JSON foradditionalModelResponseFieldPaths
.[ "/stop_sequence" ]
For information about the JSON Pointer syntax, see the Internet Engineering Task Force (IETF) documentation.
ConverseStream
rejects an empty JSON Pointer or incorrectly structured JSON Pointer with a400
error code. if the JSON Pointer is valid, but the requested field is not in the model response, it is ignored byConverseStream
. - guardrailConfig
-
- Type: GuardrailStreamConfiguration structure
Configuration information for a guardrail that you want to use in the request.
- inferenceConfig
-
- Type: InferenceConfiguration structure
Inference parameters to pass to the model.
ConverseStream
supports a base set of inference parameters. If you need to pass additional parameters that the model supports, use theadditionalModelRequestFields
request field. - messages
-
- Required: Yes
- Type: Array of Message structures
The messages that you want to send to the model.
- modelId
-
- Required: Yes
- Type: string
The ID for the model.
The
modelId
to provide depends on the type of model or throughput that you use:-
If you use a base model, specify the model ID or its ARN. For a list of model IDs for base models, see Amazon Bedrock base model IDs (on-demand throughput) in the Amazon Bedrock User Guide.
-
If you use an inference profile, specify the inference profile ID or its ARN. For a list of inference profile IDs, see Supported Regions and models for cross-region inference in the Amazon Bedrock User Guide.
-
If you use a provisioned model, specify the ARN of the Provisioned Throughput. For more information, see Run inference using a Provisioned Throughput in the Amazon Bedrock User Guide.
-
If you use a custom model, first purchase Provisioned Throughput for it. Then specify the ARN of the resulting provisioned model. For more information, see Use a custom model in Amazon Bedrock in the Amazon Bedrock User Guide.
The Converse API doesn't support imported models.
- system
-
- Type: Array of SystemContentBlock structures
A system prompt to send to the model.
- toolConfig
-
- Type: ToolConfiguration structure
Configuration information for the tools that the model can use when generating a response.
This field is only supported by Anthropic Claude 3 models.
Result Syntax
[ 'stream' => [ // EventParsingIterator 'contentBlockDelta' => [ 'contentBlockIndex' => <integer>, 'delta' => [ 'text' => '<string>', 'toolUse' => [ 'input' => '<string>', ], ], ], 'contentBlockStart' => [ 'contentBlockIndex' => <integer>, 'start' => [ 'toolUse' => [ 'name' => '<string>', 'toolUseId' => '<string>', ], ], ], 'contentBlockStop' => [ 'contentBlockIndex' => <integer>, ], 'internalServerException' => [ 'message' => '<string>', ], 'messageStart' => [ 'role' => 'user|assistant', ], 'messageStop' => [ 'additionalModelResponseFields' => [ ], 'stopReason' => 'end_turn|tool_use|max_tokens|stop_sequence|guardrail_intervened|content_filtered', ], 'metadata' => [ 'metrics' => [ 'latencyMs' => <integer>, ], 'trace' => [ 'guardrail' => [ 'inputAssessment' => [ '<String>' => [ 'contentPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED', 'confidence' => 'NONE|LOW|MEDIUM|HIGH', 'filterStrength' => 'NONE|LOW|MEDIUM|HIGH', 'type' => 'INSULTS|HATE|SEXUAL|VIOLENCE|MISCONDUCT|PROMPT_ATTACK', ], // ... ], ], 'contextualGroundingPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED|NONE', 'score' => <float>, 'threshold' => <float>, 'type' => 'GROUNDING|RELEVANCE', ], // ... ], ], 'invocationMetrics' => [ 'guardrailCoverage' => [ 'textCharacters' => [ 'guarded' => <integer>, 'total' => <integer>, ], ], 'guardrailProcessingLatency' => <integer>, 'usage' => [ 'contentPolicyUnits' => <integer>, 'contextualGroundingPolicyUnits' => <integer>, 'sensitiveInformationPolicyFreeUnits' => <integer>, 'sensitiveInformationPolicyUnits' => <integer>, 'topicPolicyUnits' => <integer>, 'wordPolicyUnits' => <integer>, ], ], 'sensitiveInformationPolicy' => [ 'piiEntities' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'type' => 'ADDRESS|AGE|AWS_ACCESS_KEY|AWS_SECRET_KEY|CA_HEALTH_NUMBER|CA_SOCIAL_INSURANCE_NUMBER|CREDIT_DEBIT_CARD_CVV|CREDIT_DEBIT_CARD_EXPIRY|CREDIT_DEBIT_CARD_NUMBER|DRIVER_ID|EMAIL|INTERNATIONAL_BANK_ACCOUNT_NUMBER|IP_ADDRESS|LICENSE_PLATE|MAC_ADDRESS|NAME|PASSWORD|PHONE|PIN|SWIFT_CODE|UK_NATIONAL_HEALTH_SERVICE_NUMBER|UK_NATIONAL_INSURANCE_NUMBER|UK_UNIQUE_TAXPAYER_REFERENCE_NUMBER|URL|USERNAME|US_BANK_ACCOUNT_NUMBER|US_BANK_ROUTING_NUMBER|US_INDIVIDUAL_TAX_IDENTIFICATION_NUMBER|US_PASSPORT_NUMBER|US_SOCIAL_SECURITY_NUMBER|VEHICLE_IDENTIFICATION_NUMBER', ], // ... ], 'regexes' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'name' => '<string>', 'regex' => '<string>', ], // ... ], ], 'topicPolicy' => [ 'topics' => [ [ 'action' => 'BLOCKED', 'name' => '<string>', 'type' => 'DENY', ], // ... ], ], 'wordPolicy' => [ 'customWords' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', ], // ... ], 'managedWordLists' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', 'type' => 'PROFANITY', ], // ... ], ], ], // ... ], 'modelOutput' => ['<string>', ...], 'outputAssessments' => [ '<String>' => [ [ 'contentPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED', 'confidence' => 'NONE|LOW|MEDIUM|HIGH', 'filterStrength' => 'NONE|LOW|MEDIUM|HIGH', 'type' => 'INSULTS|HATE|SEXUAL|VIOLENCE|MISCONDUCT|PROMPT_ATTACK', ], // ... ], ], 'contextualGroundingPolicy' => [ 'filters' => [ [ 'action' => 'BLOCKED|NONE', 'score' => <float>, 'threshold' => <float>, 'type' => 'GROUNDING|RELEVANCE', ], // ... ], ], 'invocationMetrics' => [ 'guardrailCoverage' => [ 'textCharacters' => [ 'guarded' => <integer>, 'total' => <integer>, ], ], 'guardrailProcessingLatency' => <integer>, 'usage' => [ 'contentPolicyUnits' => <integer>, 'contextualGroundingPolicyUnits' => <integer>, 'sensitiveInformationPolicyFreeUnits' => <integer>, 'sensitiveInformationPolicyUnits' => <integer>, 'topicPolicyUnits' => <integer>, 'wordPolicyUnits' => <integer>, ], ], 'sensitiveInformationPolicy' => [ 'piiEntities' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'type' => 'ADDRESS|AGE|AWS_ACCESS_KEY|AWS_SECRET_KEY|CA_HEALTH_NUMBER|CA_SOCIAL_INSURANCE_NUMBER|CREDIT_DEBIT_CARD_CVV|CREDIT_DEBIT_CARD_EXPIRY|CREDIT_DEBIT_CARD_NUMBER|DRIVER_ID|EMAIL|INTERNATIONAL_BANK_ACCOUNT_NUMBER|IP_ADDRESS|LICENSE_PLATE|MAC_ADDRESS|NAME|PASSWORD|PHONE|PIN|SWIFT_CODE|UK_NATIONAL_HEALTH_SERVICE_NUMBER|UK_NATIONAL_INSURANCE_NUMBER|UK_UNIQUE_TAXPAYER_REFERENCE_NUMBER|URL|USERNAME|US_BANK_ACCOUNT_NUMBER|US_BANK_ROUTING_NUMBER|US_INDIVIDUAL_TAX_IDENTIFICATION_NUMBER|US_PASSPORT_NUMBER|US_SOCIAL_SECURITY_NUMBER|VEHICLE_IDENTIFICATION_NUMBER', ], // ... ], 'regexes' => [ [ 'action' => 'ANONYMIZED|BLOCKED', 'match' => '<string>', 'name' => '<string>', 'regex' => '<string>', ], // ... ], ], 'topicPolicy' => [ 'topics' => [ [ 'action' => 'BLOCKED', 'name' => '<string>', 'type' => 'DENY', ], // ... ], ], 'wordPolicy' => [ 'customWords' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', ], // ... ], 'managedWordLists' => [ [ 'action' => 'BLOCKED', 'match' => '<string>', 'type' => 'PROFANITY', ], // ... ], ], ], // ... ], // ... ], ], ], 'usage' => [ 'inputTokens' => <integer>, 'outputTokens' => <integer>, 'totalTokens' => <integer>, ], ], 'modelStreamErrorException' => [ 'message' => '<string>', 'originalMessage' => '<string>', 'originalStatusCode' => <integer>, ], 'serviceUnavailableException' => [ 'message' => '<string>', ], 'throttlingException' => [ 'message' => '<string>', ], 'validationException' => [ 'message' => '<string>', ], ], ]
Result Details
Members
- stream
-
- Type: EventParsingIterator supplying the following structures: MessageStartEvent, ContentBlockStartEvent, ContentBlockDeltaEvent, ContentBlockStopEvent, MessageStopEvent, ConverseStreamMetadataEvent, InternalServerException, ModelStreamErrorException, ValidationException, ThrottlingException, ServiceUnavailableException
The output stream that the model generated.
Using an EventParsingIterator
To use an EventParsingIterator, you will need to loop over the events it will generate and check the top-level field to determine which type of event it is.
foreach($result['stream'] as $event) { if (isset($event['contentBlockDelta'])) { // Handle the 'contentBlockDelta' event. } else if (isset($event['contentBlockStart'])) { // Handle the 'contentBlockStart' event. } else if (isset($event['contentBlockStop'])) { // Handle the 'contentBlockStop' event. } else if (isset($event['internalServerException'])) { // Handle the 'internalServerException' event. } else if (isset($event['messageStart'])) { // Handle the 'messageStart' event. } else if (isset($event['messageStop'])) { // Handle the 'messageStop' event. } else if (isset($event['metadata'])) { // Handle the 'metadata' event. } else if (isset($event['modelStreamErrorException'])) { // Handle the 'modelStreamErrorException' event. } else if (isset($event['serviceUnavailableException'])) { // Handle the 'serviceUnavailableException' event. } else if (isset($event['throttlingException'])) { // Handle the 'throttlingException' event. } else if (isset($event['validationException'])) { // Handle the 'validationException' event. } }
Errors
- AccessDeniedException:
The request is denied because of missing access permissions.
- ResourceNotFoundException:
The specified resource ARN was not found. Check the ARN and try your request again.
- ThrottlingException:
Your request was throttled because of service-wide limitations. Resubmit your request later or in a different region. You can also purchase Provisioned Throughput to increase the rate or number of tokens you can process.
- ModelTimeoutException:
The request took too long to process. Processing time exceeded the model timeout length.
- InternalServerException:
An internal server error occurred. Retry your request.
- ServiceUnavailableException:
The service isn't currently available. Try again later.
- ValidationException:
Input validation failed. Check your request parameters and retry the request.
- ModelNotReadyException:
The model specified in the request is not ready to serve inference requests. The AWS SDK will automatically retry the operation up to 5 times. For information about configuring automatic retries, see Retry behavior in the AWS SDKs and Tools reference guide.
- ModelErrorException:
The request failed due to an error while processing the model.
InvokeModel
$result = $client->invokeModel
([/* ... */]); $promise = $client->invokeModelAsync
([/* ... */]);
Invokes the specified Amazon Bedrock model to run inference using the prompt and inference parameters provided in the request body. You use model inference to generate text, images, and embeddings.
For example code, see Invoke model code examples in the Amazon Bedrock User Guide.
This operation requires permission for the bedrock:InvokeModel
action.
Parameter Syntax
$result = $client->invokeModel([ 'accept' => '<string>', 'body' => <string || resource || Psr\Http\Message\StreamInterface>, // REQUIRED 'contentType' => '<string>', 'guardrailIdentifier' => '<string>', 'guardrailVersion' => '<string>', 'modelId' => '<string>', // REQUIRED 'trace' => 'ENABLED|DISABLED', ]);
Parameter Details
Members
- accept
-
- Type: string
The desired MIME type of the inference body in the response. The default value is
application/json
. - body
-
- Required: Yes
- Type: blob (string|resource|Psr\Http\Message\StreamInterface)
The prompt and inference parameters in the format specified in the
contentType
in the header. You must provide the body in JSON format. To see the format and content of the request and response bodies for different models, refer to Inference parameters. For more information, see Run inference in the Bedrock User Guide. - contentType
-
- Type: string
The MIME type of the input data in the request. You must specify
application/json
. - guardrailIdentifier
-
- Type: string
The unique identifier of the guardrail that you want to use. If you don't provide a value, no guardrail is applied to the invocation.
An error will be thrown in the following situations.
-
You don't provide a guardrail identifier but you specify the
amazon-bedrock-guardrailConfig
field in the request body. -
You enable the guardrail but the
contentType
isn'tapplication/json
. -
You provide a guardrail identifier, but
guardrailVersion
isn't specified.
- guardrailVersion
-
- Type: string
The version number for the guardrail. The value can also be
DRAFT
. - modelId
-
- Required: Yes
- Type: string
The unique identifier of the model to invoke to run inference.
The
modelId
to provide depends on the type of model that you use:-
If you use a base model, specify the model ID or its ARN. For a list of model IDs for base models, see Amazon Bedrock base model IDs (on-demand throughput) in the Amazon Bedrock User Guide.
-
If you use a provisioned model, specify the ARN of the Provisioned Throughput. For more information, see Run inference using a Provisioned Throughput in the Amazon Bedrock User Guide.
-
If you use a custom model, first purchase Provisioned Throughput for it. Then specify the ARN of the resulting provisioned model. For more information, see Use a custom model in Amazon Bedrock in the Amazon Bedrock User Guide.
-
If you use an imported model, specify the ARN of the imported model. You can get the model ARN from a successful call to CreateModelImportJob or from the Imported models page in the Amazon Bedrock console.
- trace
-
- Type: string
Specifies whether to enable or disable the Bedrock trace. If enabled, you can see the full Bedrock trace.
Result Syntax
[ 'body' => <string || resource || Psr\Http\Message\StreamInterface>, 'contentType' => '<string>', ]
Result Details
Members
- body
-
- Required: Yes
- Type: blob (string|resource|Psr\Http\Message\StreamInterface)
Inference response from the model in the format specified in the
contentType
header. To see the format and content of the request and response bodies for different models, refer to Inference parameters. - contentType
-
- Required: Yes
- Type: string
The MIME type of the inference result.
Errors
- AccessDeniedException:
The request is denied because of missing access permissions.
- ResourceNotFoundException:
The specified resource ARN was not found. Check the ARN and try your request again.
- ThrottlingException:
Your request was throttled because of service-wide limitations. Resubmit your request later or in a different region. You can also purchase Provisioned Throughput to increase the rate or number of tokens you can process.
- ModelTimeoutException:
The request took too long to process. Processing time exceeded the model timeout length.
- InternalServerException:
An internal server error occurred. Retry your request.
- ServiceUnavailableException:
The service isn't currently available. Try again later.
- ValidationException:
Input validation failed. Check your request parameters and retry the request.
- ModelNotReadyException:
The model specified in the request is not ready to serve inference requests. The AWS SDK will automatically retry the operation up to 5 times. For information about configuring automatic retries, see Retry behavior in the AWS SDKs and Tools reference guide.
- ServiceQuotaExceededException:
Your request exceeds the service quota for your account. You can view your quotas at Viewing service quotas. You can resubmit your request later.
- ModelErrorException:
The request failed due to an error while processing the model.
InvokeModelWithResponseStream
$result = $client->invokeModelWithResponseStream
([/* ... */]); $promise = $client->invokeModelWithResponseStreamAsync
([/* ... */]);
Invoke the specified Amazon Bedrock model to run inference using the prompt and inference parameters provided in the request body. The response is returned in a stream.
To see if a model supports streaming, call GetFoundationModel and check the responseStreamingSupported
field in the response.
The CLI doesn't support streaming operations in Amazon Bedrock, including InvokeModelWithResponseStream
.
For example code, see Invoke model with streaming code example in the Amazon Bedrock User Guide.
This operation requires permissions to perform the bedrock:InvokeModelWithResponseStream
action.
Parameter Syntax
$result = $client->invokeModelWithResponseStream([ 'accept' => '<string>', 'body' => <string || resource || Psr\Http\Message\StreamInterface>, // REQUIRED 'contentType' => '<string>', 'guardrailIdentifier' => '<string>', 'guardrailVersion' => '<string>', 'modelId' => '<string>', // REQUIRED 'trace' => 'ENABLED|DISABLED', ]);
Parameter Details
Members
- accept
-
- Type: string
The desired MIME type of the inference body in the response. The default value is
application/json
. - body
-
- Required: Yes
- Type: blob (string|resource|Psr\Http\Message\StreamInterface)
The prompt and inference parameters in the format specified in the
contentType
in the header. You must provide the body in JSON format. To see the format and content of the request and response bodies for different models, refer to Inference parameters. For more information, see Run inference in the Bedrock User Guide. - contentType
-
- Type: string
The MIME type of the input data in the request. You must specify
application/json
. - guardrailIdentifier
-
- Type: string
The unique identifier of the guardrail that you want to use. If you don't provide a value, no guardrail is applied to the invocation.
An error is thrown in the following situations.
-
You don't provide a guardrail identifier but you specify the
amazon-bedrock-guardrailConfig
field in the request body. -
You enable the guardrail but the
contentType
isn'tapplication/json
. -
You provide a guardrail identifier, but
guardrailVersion
isn't specified.
- guardrailVersion
-
- Type: string
The version number for the guardrail. The value can also be
DRAFT
. - modelId
-
- Required: Yes
- Type: string
The unique identifier of the model to invoke to run inference.
The
modelId
to provide depends on the type of model that you use:-
If you use a base model, specify the model ID or its ARN. For a list of model IDs for base models, see Amazon Bedrock base model IDs (on-demand throughput) in the Amazon Bedrock User Guide.
-
If you use a provisioned model, specify the ARN of the Provisioned Throughput. For more information, see Run inference using a Provisioned Throughput in the Amazon Bedrock User Guide.
-
If you use a custom model, first purchase Provisioned Throughput for it. Then specify the ARN of the resulting provisioned model. For more information, see Use a custom model in Amazon Bedrock in the Amazon Bedrock User Guide.
-
If you use an imported model, specify the ARN of the imported model. You can get the model ARN from a successful call to CreateModelImportJob or from the Imported models page in the Amazon Bedrock console.
- trace
-
- Type: string
Specifies whether to enable or disable the Bedrock trace. If enabled, you can see the full Bedrock trace.
Result Syntax
[ 'body' => [ // EventParsingIterator 'chunk' => [ 'bytes' => <string || resource || Psr\Http\Message\StreamInterface>, ], 'internalServerException' => [ 'message' => '<string>', ], 'modelStreamErrorException' => [ 'message' => '<string>', 'originalMessage' => '<string>', 'originalStatusCode' => <integer>, ], 'modelTimeoutException' => [ 'message' => '<string>', ], 'serviceUnavailableException' => [ 'message' => '<string>', ], 'throttlingException' => [ 'message' => '<string>', ], 'validationException' => [ 'message' => '<string>', ], ], 'contentType' => '<string>', ]
Result Details
Members
- body
-
- Required: Yes
- Type: EventParsingIterator supplying the following structures: PayloadPart, InternalServerException, ModelStreamErrorException, ValidationException, ThrottlingException, ModelTimeoutException, ServiceUnavailableException
Inference response from the model in the format specified by the
contentType
header. To see the format and content of this field for different models, refer to Inference parameters. - contentType
-
- Required: Yes
- Type: string
The MIME type of the inference result.
Using an EventParsingIterator
To use an EventParsingIterator, you will need to loop over the events it will generate and check the top-level field to determine which type of event it is.
foreach($result['body'] as $event) { if (isset($event['chunk'])) { // Handle the 'chunk' event. } else if (isset($event['internalServerException'])) { // Handle the 'internalServerException' event. } else if (isset($event['modelStreamErrorException'])) { // Handle the 'modelStreamErrorException' event. } else if (isset($event['modelTimeoutException'])) { // Handle the 'modelTimeoutException' event. } else if (isset($event['serviceUnavailableException'])) { // Handle the 'serviceUnavailableException' event. } else if (isset($event['throttlingException'])) { // Handle the 'throttlingException' event. } else if (isset($event['validationException'])) { // Handle the 'validationException' event. } }
Errors
- AccessDeniedException:
The request is denied because of missing access permissions.
- ResourceNotFoundException:
The specified resource ARN was not found. Check the ARN and try your request again.
- ThrottlingException:
Your request was throttled because of service-wide limitations. Resubmit your request later or in a different region. You can also purchase Provisioned Throughput to increase the rate or number of tokens you can process.
- ModelTimeoutException:
The request took too long to process. Processing time exceeded the model timeout length.
- InternalServerException:
An internal server error occurred. Retry your request.
- ServiceUnavailableException:
The service isn't currently available. Try again later.
- ModelStreamErrorException:
An error occurred while streaming the response. Retry your request.
- ValidationException:
Input validation failed. Check your request parameters and retry the request.
- ModelNotReadyException:
The model specified in the request is not ready to serve inference requests. The AWS SDK will automatically retry the operation up to 5 times. For information about configuring automatic retries, see Retry behavior in the AWS SDKs and Tools reference guide.
- ServiceQuotaExceededException:
Your request exceeds the service quota for your account. You can view your quotas at Viewing service quotas. You can resubmit your request later.
- ModelErrorException:
The request failed due to an error while processing the model.
Shapes
AccessDeniedException
Description
The request is denied because of missing access permissions.
Members
- message
-
- Type: string
AnyToolChoice
Description
The model must request at least one tool (no text is generated). For example, {"any" : {}}
.
Members
AutoToolChoice
Description
The Model automatically decides if a tool should be called or whether to generate text instead. For example, {"auto" : {}}
.
Members
ContentBlock
Description
A block of content for a message that you pass to, or receive from, a model with the Converse or ConverseStream API operations.
Members
- document
-
- Type: DocumentBlock structure
A document to include in the message.
- guardContent
-
- Type: GuardrailConverseContentBlock structure
Contains the content to assess with the guardrail. If you don't specify
guardContent
in a call to the Converse API, the guardrail (if passed in the Converse API) assesses the entire message.For more information, see Use a guardrail with the Converse API in the Amazon Bedrock User Guide.
</p>
- image
-
- Type: ImageBlock structure
Image to include in the message.
This field is only supported by Anthropic Claude 3 models.
- text
-
- Type: string
Text to include in the message.
- toolResult
-
- Type: ToolResultBlock structure
The result for a tool request that a model makes.
- toolUse
-
- Type: ToolUseBlock structure
Information about a tool use request from a model.
ContentBlockDelta
Description
A bock of content in a streaming response.
Members
- text
-
- Type: string
The content text.
- toolUse
-
- Type: ToolUseBlockDelta structure
Information about a tool that the model is requesting to use.
ContentBlockDeltaEvent
Description
The content block delta event.
Members
- contentBlockIndex
-
- Required: Yes
- Type: int
The block index for a content block delta event.
- delta
-
- Required: Yes
- Type: ContentBlockDelta structure
The delta for a content block delta event.
ContentBlockStart
Description
Content block start information.
Members
- toolUse
-
- Type: ToolUseBlockStart structure
Information about a tool that the model is requesting to use.
ContentBlockStartEvent
Description
Content block start event.
Members
- contentBlockIndex
-
- Required: Yes
- Type: int
The index for a content block start event.
- start
-
- Required: Yes
- Type: ContentBlockStart structure
Start information about a content block start event.
ContentBlockStopEvent
Description
A content block stop event.
Members
- contentBlockIndex
-
- Required: Yes
- Type: int
The index for a content block.
ConverseMetrics
Description
Metrics for a call to Converse.
Members
- latencyMs
-
- Required: Yes
- Type: long (int|float)
The latency of the call to
Converse
, in milliseconds.
ConverseOutput
ConverseStreamMetadataEvent
Description
A conversation stream metadata event.
Members
- metrics
-
- Required: Yes
- Type: ConverseStreamMetrics structure
The metrics for the conversation stream metadata event.
- trace
-
- Type: ConverseStreamTrace structure
The trace object in the response from ConverseStream that contains information about the guardrail behavior.
- usage
-
- Required: Yes
- Type: TokenUsage structure
Usage information for the conversation stream event.
ConverseStreamMetrics
Description
Metrics for the stream.
Members
- latencyMs
-
- Required: Yes
- Type: long (int|float)
The latency for the streaming request, in milliseconds.
ConverseStreamOutput
Description
The messages output stream
Members
- contentBlockDelta
-
- Type: ContentBlockDeltaEvent structure
The messages output content block delta.
- contentBlockStart
-
- Type: ContentBlockStartEvent structure
Start information for a content block.
- contentBlockStop
-
- Type: ContentBlockStopEvent structure
Stop information for a content block.
- internalServerException
-
- Type: InternalServerException structure
An internal server error occurred. Retry your request.
- messageStart
-
- Type: MessageStartEvent structure
Message start information.
- messageStop
-
- Type: MessageStopEvent structure
Message stop information.
- metadata
-
- Type: ConverseStreamMetadataEvent structure
Metadata for the converse output stream.
- modelStreamErrorException
-
- Type: ModelStreamErrorException structure
A streaming error occurred. Retry your request.
- serviceUnavailableException
-
- Type: ServiceUnavailableException structure
The service isn't currently available. Try again later.
- throttlingException
-
- Type: ThrottlingException structure
The number of requests exceeds the limit. Resubmit your request later.
- validationException
-
- Type: ValidationException structure
Input validation failed. Check your request parameters and retry the request.
ConverseStreamTrace
Description
The trace object in a response from ConverseStream. Currently, you can only trace guardrails.
Members
- guardrail
-
- Type: GuardrailTraceAssessment structure
The guardrail trace object.
ConverseTrace
Description
The trace object in a response from Converse. Currently, you can only trace guardrails.
Members
- guardrail
-
- Type: GuardrailTraceAssessment structure
The guardrail trace object.
Document
Members
DocumentBlock
Description
A document to include in a message.
Members
- format
-
- Required: Yes
- Type: string
The format of a document, or its extension.
- name
-
- Required: Yes
- Type: string
A name for the document. The name can only contain the following characters:
-
Alphanumeric characters
-
Whitespace characters (no more than one in a row)
-
Hyphens
-
Parentheses
-
Square brackets
This field is vulnerable to prompt injections, because the model might inadvertently interpret it as instructions. Therefore, we recommend that you specify a neutral name.
- source
-
- Required: Yes
- Type: DocumentSource structure
Contains the content of the document.
DocumentSource
Description
Contains the content of a document.
Members
- bytes
-
- Type: blob (string|resource|Psr\Http\Message\StreamInterface)
The raw bytes for the document. If you use an Amazon Web Services SDK, you don't need to encode the bytes in base64.
GuardrailAssessment
Description
A behavior assessment of the guardrail policies used in a call to the Converse API.
Members
- contentPolicy
-
- Type: GuardrailContentPolicyAssessment structure
The content policy.
- contextualGroundingPolicy
-
- Type: GuardrailContextualGroundingPolicyAssessment structure
The contextual grounding policy used for the guardrail assessment.
- invocationMetrics
-
- Type: GuardrailInvocationMetrics structure
The invocation metrics for the guardrail assessment.
- sensitiveInformationPolicy
-
- Type: GuardrailSensitiveInformationPolicyAssessment structure
The sensitive information policy.
- topicPolicy
-
- Type: GuardrailTopicPolicyAssessment structure
The topic policy.
- wordPolicy
-
- Type: GuardrailWordPolicyAssessment structure
The word policy.
GuardrailConfiguration
Description
Configuration information for a guardrail that you use with the Converse operation.
Members
- guardrailIdentifier
-
- Required: Yes
- Type: string
The identifier for the guardrail.
- guardrailVersion
-
- Required: Yes
- Type: string
The version of the guardrail.
- trace
-
- Type: string
The trace behavior for the guardrail.
GuardrailContentBlock
Description
The content block to be evaluated by the guardrail.
Members
- text
-
- Type: GuardrailTextBlock structure
Text within content block to be evaluated by the guardrail.
GuardrailContentFilter
Description
The content filter for a guardrail.
Members
- action
-
- Required: Yes
- Type: string
The guardrail action.
- confidence
-
- Required: Yes
- Type: string
The guardrail confidence.
- filterStrength
-
- Type: string
The filter strength setting for the guardrail content filter.
- type
-
- Required: Yes
- Type: string
The guardrail type.
GuardrailContentPolicyAssessment
Description
An assessment of a content policy for a guardrail.
Members
- filters
-
- Required: Yes
- Type: Array of GuardrailContentFilter structures
The content policy filters.
GuardrailContextualGroundingFilter
Description
The details for the guardrails contextual grounding filter.
Members
- action
-
- Required: Yes
- Type: string
The action performed by the guardrails contextual grounding filter.
- score
-
- Required: Yes
- Type: double
The score generated by contextual grounding filter.
- threshold
-
- Required: Yes
- Type: double
The threshold used by contextual grounding filter to determine whether the content is grounded or not.
- type
-
- Required: Yes
- Type: string
The contextual grounding filter type.
GuardrailContextualGroundingPolicyAssessment
Description
The policy assessment details for the guardrails contextual grounding filter.
Members
- filters
-
- Type: Array of GuardrailContextualGroundingFilter structures
The filter details for the guardrails contextual grounding filter.
GuardrailConverseContentBlock
Description
A content block for selective guarding with the Converse or ConverseStream API operations.
Members
- text
-
- Type: GuardrailConverseTextBlock structure
The text to guard.
GuardrailConverseTextBlock
Description
A text block that contains text that you want to assess with a guardrail. For more information, see GuardrailConverseContentBlock.
Members
- qualifiers
-
- Type: Array of strings
The qualifier details for the guardrails contextual grounding filter.
- text
-
- Required: Yes
- Type: string
The text that you want to guard.
GuardrailCoverage
Description
The action of the guardrail coverage details.
Members
- textCharacters
-
- Type: GuardrailTextCharactersCoverage structure
The text characters of the guardrail coverage details.
GuardrailCustomWord
Description
A custom word configured in a guardrail.
Members
- action
-
- Required: Yes
- Type: string
The action for the custom word.
- match
-
- Required: Yes
- Type: string
The match for the custom word.
GuardrailInvocationMetrics
Description
The invocation metrics for the guardrail.
Members
- guardrailCoverage
-
- Type: GuardrailCoverage structure
The coverage details for the guardrail invocation metrics.
- guardrailProcessingLatency
-
- Type: long (int|float)
The processing latency details for the guardrail invocation metrics.
- usage
-
- Type: GuardrailUsage structure
The usage details for the guardrail invocation metrics.
GuardrailManagedWord
Description
A managed word configured in a guardrail.
Members
- action
-
- Required: Yes
- Type: string
The action for the managed word.
- match
-
- Required: Yes
- Type: string
The match for the managed word.
- type
-
- Required: Yes
- Type: string
The type for the managed word.
GuardrailOutputContent
Description
The output content produced by the guardrail.
Members
- text
-
- Type: string
The specific text for the output content produced by the guardrail.
GuardrailPiiEntityFilter
Description
A Personally Identifiable Information (PII) entity configured in a guardrail.
Members
- action
-
- Required: Yes
- Type: string
The PII entity filter action.
- match
-
- Required: Yes
- Type: string
The PII entity filter match.
- type
-
- Required: Yes
- Type: string
The PII entity filter type.
GuardrailRegexFilter
Description
A Regex filter configured in a guardrail.
Members
- action
-
- Required: Yes
- Type: string
The region filter action.
- match
-
- Type: string
The regesx filter match.
- name
-
- Type: string
The regex filter name.
- regex
-
- Type: string
The regex query.
GuardrailSensitiveInformationPolicyAssessment
Description
The assessment for aPersonally Identifiable Information (PII) policy.
Members
- piiEntities
-
- Required: Yes
- Type: Array of GuardrailPiiEntityFilter structures
The PII entities in the assessment.
- regexes
-
- Required: Yes
- Type: Array of GuardrailRegexFilter structures
The regex queries in the assessment.
GuardrailStreamConfiguration
Description
Configuration information for a guardrail that you use with the ConverseStream action.
Members
- guardrailIdentifier
-
- Required: Yes
- Type: string
The identifier for the guardrail.
- guardrailVersion
-
- Required: Yes
- Type: string
The version of the guardrail.
- streamProcessingMode
-
- Type: string
The processing mode.
The processing mode. For more information, see Configure streaming response behavior in the Amazon Bedrock User Guide.
- trace
-
- Type: string
The trace behavior for the guardrail.
GuardrailTextBlock
Description
The text block to be evaluated by the guardrail.
Members
- qualifiers
-
- Type: Array of strings
The qualifiers describing the text block.
- text
-
- Required: Yes
- Type: string
The input text details to be evaluated by the guardrail.
GuardrailTextCharactersCoverage
Description
The guardrail coverage for the text characters.
Members
- guarded
-
- Type: int
The text characters that were guarded by the guardrail coverage.
- total
-
- Type: int
The total text characters by the guardrail coverage.
GuardrailTopic
Description
Information about a topic guardrail.
Members
- action
-
- Required: Yes
- Type: string
The action the guardrail should take when it intervenes on a topic.
- name
-
- Required: Yes
- Type: string
The name for the guardrail.
- type
-
- Required: Yes
- Type: string
The type behavior that the guardrail should perform when the model detects the topic.
GuardrailTopicPolicyAssessment
Description
A behavior assessment of a topic policy.
Members
- topics
-
- Required: Yes
- Type: Array of GuardrailTopic structures
The topics in the assessment.
GuardrailTraceAssessment
Description
A Top level guardrail trace object. For more information, see ConverseTrace.
Members
- inputAssessment
-
- Type: Associative array of custom strings keys (String) to GuardrailAssessment structures
The input assessment.
- modelOutput
-
- Type: Array of strings
The output from the model.
- outputAssessments
-
- Type: Associative array of custom strings keys (String) to GuardrailAssessment structuress
the output assessments.
GuardrailUsage
Description
The details on the use of the guardrail.
Members
- contentPolicyUnits
-
- Required: Yes
- Type: int
The content policy units processed by the guardrail.
- contextualGroundingPolicyUnits
-
- Required: Yes
- Type: int
The contextual grounding policy units processed by the guardrail.
- sensitiveInformationPolicyFreeUnits
-
- Required: Yes
- Type: int
The sensitive information policy free units processed by the guardrail.
- sensitiveInformationPolicyUnits
-
- Required: Yes
- Type: int
The sensitive information policy units processed by the guardrail.
- topicPolicyUnits
-
- Required: Yes
- Type: int
The topic policy units processed by the guardrail.
- wordPolicyUnits
-
- Required: Yes
- Type: int
The word policy units processed by the guardrail.
GuardrailWordPolicyAssessment
Description
The word policy assessment.
Members
- customWords
-
- Required: Yes
- Type: Array of GuardrailCustomWord structures
Custom words in the assessment.
- managedWordLists
-
- Required: Yes
- Type: Array of GuardrailManagedWord structures
Managed word lists in the assessment.
ImageBlock
Description
Image content for a message.
Members
- format
-
- Required: Yes
- Type: string
The format of the image.
- source
-
- Required: Yes
- Type: ImageSource structure
The source for the image.
ImageSource
Description
The source for an image.
Members
- bytes
-
- Type: blob (string|resource|Psr\Http\Message\StreamInterface)
The raw image bytes for the image. If you use an AWS SDK, you don't need to encode the image bytes in base64.
InferenceConfiguration
Description
Base inference parameters to pass to a model in a call to Converse or ConverseStream. For more information, see Inference parameters for foundation models.
If you need to pass additional parameters that the model supports, use the additionalModelRequestFields
request field in the call to Converse
or ConverseStream
. For more information, see Model parameters.
Members
- maxTokens
-
- Type: int
The maximum number of tokens to allow in the generated response. The default value is the maximum allowed value for the model that you are using. For more information, see Inference parameters for foundation models.
- stopSequences
-
- Type: Array of strings
A list of stop sequences. A stop sequence is a sequence of characters that causes the model to stop generating the response.
- temperature
-
- Type: float
The likelihood of the model selecting higher-probability options while generating a response. A lower value makes the model more likely to choose higher-probability options, while a higher value makes the model more likely to choose lower-probability options.
The default value is the default value for the model that you are using. For more information, see Inference parameters for foundation models.
- topP
-
- Type: float
The percentage of most-likely candidates that the model considers for the next token. For example, if you choose a value of 0.8 for
topP
, the model selects from the top 80% of the probability distribution of tokens that could be next in the sequence.The default value is the default value for the model that you are using. For more information, see Inference parameters for foundation models.
InternalServerException
Description
An internal server error occurred. Retry your request.
Members
- message
-
- Type: string
Message
Description
A message input, or returned from, a call to Converse or ConverseStream.
Members
- content
-
- Required: Yes
- Type: Array of ContentBlock structures
The message content. Note the following restrictions:
-
You can include up to 20 images. Each image's size, height, and width must be no more than 3.75 MB, 8000 px, and 8000 px, respectively.
-
You can include up to five documents. Each document's size must be no more than 4.5 MB.
-
If you include a
ContentBlock
with adocument
field in the array, you must also include aContentBlock
with atext
field. -
You can only include images and documents if the
role
isuser
.
- role
-
- Required: Yes
- Type: string
The role that the message plays in the message.
MessageStartEvent
Description
The start of a message.
Members
- role
-
- Required: Yes
- Type: string
The role for the message.
MessageStopEvent
Description
The stop event for a message.
Members
- additionalModelResponseFields
-
- Type: document (null|bool|string|numeric) or an (array|associative array) whose members are all valid documents
The additional model response fields.
- stopReason
-
- Required: Yes
- Type: string
The reason why the model stopped generating output.
ModelErrorException
Description
The request failed due to an error while processing the model.
Members
- message
-
- Type: string
- originalStatusCode
-
- Type: int
The original status code.
- resourceName
-
- Type: string
The resource name.
ModelNotReadyException
Description
The model specified in the request is not ready to serve inference requests. The AWS SDK will automatically retry the operation up to 5 times. For information about configuring automatic retries, see Retry behavior in the AWS SDKs and Tools reference guide.
Members
- message
-
- Type: string
ModelStreamErrorException
Description
An error occurred while streaming the response. Retry your request.
Members
- message
-
- Type: string
- originalMessage
-
- Type: string
The original message.
- originalStatusCode
-
- Type: int
The original status code.
ModelTimeoutException
Description
The request took too long to process. Processing time exceeded the model timeout length.
Members
- message
-
- Type: string
PayloadPart
Description
Payload content included in the response.
Members
- bytes
-
- Type: blob (string|resource|Psr\Http\Message\StreamInterface)
Base64-encoded bytes of payload data.
ResourceNotFoundException
Description
The specified resource ARN was not found. Check the ARN and try your request again.
Members
- message
-
- Type: string
ResponseStream
Description
Definition of content in the response stream.
Members
- chunk
-
- Type: PayloadPart structure
Content included in the response.
- internalServerException
-
- Type: InternalServerException structure
An internal server error occurred. Retry your request.
- modelStreamErrorException
-
- Type: ModelStreamErrorException structure
An error occurred while streaming the response. Retry your request.
- modelTimeoutException
-
- Type: ModelTimeoutException structure
The request took too long to process. Processing time exceeded the model timeout length.
- serviceUnavailableException
-
- Type: ServiceUnavailableException structure
The service isn't currently available. Try again later.
- throttlingException
-
- Type: ThrottlingException structure
Your request was throttled because of service-wide limitations. Resubmit your request later or in a different region. You can also purchase Provisioned Throughput to increase the rate or number of tokens you can process.
- validationException
-
- Type: ValidationException structure
Input validation failed. Check your request parameters and retry the request.
ServiceQuotaExceededException
Description
Your request exceeds the service quota for your account. You can view your quotas at Viewing service quotas. You can resubmit your request later.
Members
- message
-
- Type: string
ServiceUnavailableException
Description
The service isn't currently available. Try again later.
Members
- message
-
- Type: string
SpecificToolChoice
Description
The model must request a specific tool. For example, {"tool" : {"name" : "Your tool name"}}
.
This field is only supported by Anthropic Claude 3 models.
Members
- name
-
- Required: Yes
- Type: string
The name of the tool that the model must request.
SystemContentBlock
Description
A system content block.
Members
- guardContent
-
- Type: GuardrailConverseContentBlock structure
A content block to assess with the guardrail. Use with the Converse or ConverseStream API operations.
For more information, see Use a guardrail with the Converse API in the Amazon Bedrock User Guide.
- text
-
- Type: string
A system prompt for the model.
ThrottlingException
Description
Your request was throttled because of service-wide limitations. Resubmit your request later or in a different region. You can also purchase Provisioned Throughput to increase the rate or number of tokens you can process.
Members
- message
-
- Type: string
TokenUsage
Description
The tokens used in a message API inference call.
Members
- inputTokens
-
- Required: Yes
- Type: int
The number of tokens sent in the request to the model.
- outputTokens
-
- Required: Yes
- Type: int
The number of tokens that the model generated for the request.
- totalTokens
-
- Required: Yes
- Type: int
The total of input tokens and tokens generated by the model.
Tool
Description
Information about a tool that you can use with the Converse API. For more information, see Tool use (function calling) in the Amazon Bedrock User Guide.
Members
- toolSpec
-
- Type: ToolSpecification structure
The specfication for the tool.
ToolChoice
Description
Determines which tools the model should request in a call to Converse
or ConverseStream
. ToolChoice
is only supported by Anthropic Claude 3 models and by Mistral AI Mistral Large.
Members
- any
-
- Type: AnyToolChoice structure
The model must request at least one tool (no text is generated).
- auto
-
- Type: AutoToolChoice structure
(Default). The Model automatically decides if a tool should be called or whether to generate text instead.
- tool
-
- Type: SpecificToolChoice structure
The Model must request the specified tool. Only supported by Anthropic Claude 3 models.
ToolConfiguration
Description
Configuration information for the tools that you pass to a model. For more information, see Tool use (function calling) in the Amazon Bedrock User Guide.
This field is only supported by Anthropic Claude 3, Cohere Command R, Cohere Command R+, and Mistral Large models.
Members
- toolChoice
-
- Type: ToolChoice structure
If supported by model, forces the model to request a tool.
- tools
-
- Required: Yes
- Type: Array of Tool structures
An array of tools that you want to pass to a model.
ToolInputSchema
Description
The schema for the tool. The top level schema type must be object
.
Members
- json
-
- Type: document (null|bool|string|numeric) or an (array|associative array) whose members are all valid documents
The JSON schema for the tool. For more information, see JSON Schema Reference.
ToolResultBlock
Description
A tool result block that contains the results for a tool request that the model previously made.
Members
- content
-
- Required: Yes
- Type: Array of ToolResultContentBlock structures
The content for tool result content block.
- status
-
- Type: string
The status for the tool result content block.
This field is only supported Anthropic Claude 3 models.
- toolUseId
-
- Required: Yes
- Type: string
The ID of the tool request that this is the result for.
ToolResultContentBlock
Description
The tool result content block.
Members
- document
-
- Type: DocumentBlock structure
A tool result that is a document.
- image
-
- Type: ImageBlock structure
A tool result that is an image.
This field is only supported by Anthropic Claude 3 models.
- json
-
- Type: document (null|bool|string|numeric) or an (array|associative array) whose members are all valid documents
A tool result that is JSON format data.
- text
-
- Type: string
A tool result that is text.
ToolSpecification
Description
The specification for the tool.
Members
- description
-
- Type: string
The description for the tool.
- inputSchema
-
- Required: Yes
- Type: ToolInputSchema structure
The input schema for the tool in JSON format.
- name
-
- Required: Yes
- Type: string
The name for the tool.
ToolUseBlock
Description
A tool use content block. Contains information about a tool that the model is requesting be run., The model uses the result from the tool to generate a response.
Members
- input
-
- Required: Yes
- Type: document (null|bool|string|numeric) or an (array|associative array) whose members are all valid documents
The input to pass to the tool.
- name
-
- Required: Yes
- Type: string
The name of the tool that the model wants to use.
- toolUseId
-
- Required: Yes
- Type: string
The ID for the tool request.
ToolUseBlockDelta
Description
The delta for a tool use block.
Members
- input
-
- Required: Yes
- Type: string
The input for a requested tool.
ToolUseBlockStart
Description
The start of a tool use block.
Members
- name
-
- Required: Yes
- Type: string
The name of the tool that the model is requesting to use.
- toolUseId
-
- Required: Yes
- Type: string
The ID for the tool request.
ValidationException
Description
Input validation failed. Check your request parameters and retry the request.
Members
- message
-
- Type: string