AWS ConverseStream
CertifiedStream a conversation with an Amazon Bedrock foundation model
AWS ConverseStream
Stream a conversation with an Amazon Bedrock foundation model
Streaming variant of the Converse task. Accumulates all streamed tokens and returns the complete assistant message once generation finishes. Use this when you want lower time-to-first-token latency for long responses.
type: io.kestra.plugin.aws.bedrock.ConverseStreamExamples
Stream a response from Claude on Bedrock and log the accumulated reply
id: bedrock_converse_stream
namespace: company.team
inputs:
- id: user_message
type: STRING
tasks:
- id: stream_chat
type: io.kestra.plugin.aws.bedrock.ConverseStream
region: "{{ secret('AWS_REGION') }}"
accessKeyId: "{{ secret('AWS_ACCESS_KEY_ID') }}"
secretKeyId: "{{ secret('AWS_SECRET_ACCESS_KEY') }}"
modelId: "anthropic.claude-3-5-sonnet-20241022-v2:0"
messages:
- role: user
content: "{{ inputs.user_message }}"
system: "You are a concise data engineering assistant."
inferenceConfig:
maxTokens: 2048
- id: log_reply
type: io.kestra.plugin.core.log.Log
message: "{{ outputs.stream_chat.content }}"
Properties
messages *Requiredarray
1Messages
Ordered list of conversation turns. Each entry must have a role (user or assistant) and content (text string).
modelId *Requiredstring
Model ID
The Bedrock model identifier, e.g. anthropic.claude-3-5-sonnet-20241022-v2: 0.
accessKeyId string
Access Key Id in order to connect to AWS
If no credentials are defined, we will use the default credentials provider chain to fetch credentials.
compatibilityMode booleanstring
Enable compatibility mode
Use it to connect to S3 bucket with S3 compatible services that don't support the new transport client.
endpointOverride string
The endpoint with which the SDK should communicate
This property allows you to use a different S3 compatible storage backend.
forcePathStyle booleanstring
Force path style access
Must only be used when compatibilityMode is enabled.
inferenceConfig object
Inference configuration
Optional inference parameters. Supported keys: maxTokens (integer), temperature (float 0–1), topP (float 0–1), stopSequences (list of strings).
pluginDefaultsRef Non-dynamicstring
Reference (ref) of the pluginDefaults to apply to this task.
region string
AWS region with which the SDK should communicate
secretKeyId string
Secret Key Id in order to connect to AWS
If no credentials are defined, we will use the default credentials provider chain to fetch credentials.
sessionToken string
AWS session token, retrieved from an AWS token service, used for authenticating that this user has received temporary permissions to access a given resource
If no credentials are defined, we will use the default credentials provider chain to fetch credentials.
stsEndpointOverride string
The AWS STS endpoint with which the SDKClient should communicate
stsRoleArn string
AWS STS Role
The Amazon Resource Name (ARN) of the role to assume. If set the task will use the StsAssumeRoleCredentialsProvider. If no credentials are defined, we will use the default credentials provider chain to fetch credentials.
stsRoleExternalId string
AWS STS External Id
A unique identifier that might be required when you assume a role in another account. This property is only used when an stsRoleArn is defined.
stsRoleSessionDuration string
PT15MAWS STS Session duration
The duration of the role session (default: 15 minutes, i.e., PT15M). This property is only used when an stsRoleArn is defined.
stsRoleSessionName string
AWS STS Session name
This property is only used when an stsRoleArn is defined.
system string
System prompt
Optional system-level instruction prepended before the conversation.
Outputs
content string
Content
The fully accumulated assistant reply.
inputTokens integer
Input tokens
Number of input tokens consumed.
modelId string
Model ID
The model that was invoked.
outputTokens integer
Output tokens
Number of output tokens generated.
stopReason string
Stop reason
Why the model stopped generating tokens.
