Amazon Bedrock Runtime

Technical Preview

Overview

Client-mode REST API connectors for Amazon Bedrock: triage alarms and summarize shift logs with the Converse API, create Titan embeddings with InvokeModel, count tokens, screen text with a guardrail, and list foundation models and inference profiles, all with a Bedrock API key. This is one of the templates in the Amazon Bedrock REST API app (amazon-bedrock). Deploy it alongside the other connectors of that app when they depend on each other (for example a token connector plus an API connector).

Not yet verified against the live vendor API: No Amazon Bedrock API key (or IAM credentials for SigV4) on the factory box. Needs an AWS account with Bedrock model access enabled in the chosen region (e.g. eu-west-1: Amazon Nova Lite via eu. inference profile + Titan Text Embeddings V2), a Bedrock API key (long-term ABSK… or short-term), and optionally a guardrail id for ApplyGuardrail. Then run BEDROCK_API_KEY=… python3 smoke_bedrock.py.

Auth

Amazon Bedrock API key as Authorization: Bearer {value.json.api_key} (long-term ABSK…, short-term bedrock-api-key-…). Hub secret suggested AWS_BEARER_TOKEN_BEDROCK + GET SECRET. Alternative: AWS SigV4 (service bedrock) computed outside the broker, documented in comments. Never commit keys. Auth headers used by the template: Authorization: Bearer {value.json.api_key}.

Template parameters

Placeholder	Type	Description	Required
`RouteName`	string	Name of the Amazon Bedrock Runtime connector	no
`BaseUrl`	string	Amazon Bedrock Runtime base URL without trailing slash (e.g. https://bedrock-runtime.eu-west-1.amazonaws.com)	yes

BASE_ADDRESS default: https://bedrock-runtime.eu-west-1.amazonaws.com.

Events

Event	Method	Endpoint	Request topic	Response topic	Payload fields
`Converse`	POST	`/model/{value.json.model_id}/converse`	`ai/bedrock/converse/send`	`ai/bedrock/converse/send/response`	`api_key`, `body`, `model_id`
`InvokeModel`	POST	`/model/{value.json.model_id}/invoke`	`ai/bedrock/model/invoke`	`ai/bedrock/model/invoke/response`	`api_key`, `body`, `model_id`
`CountTokens`	POST	`/model/{value.json.model_id}/count-tokens`	`ai/bedrock/tokens/count`	`ai/bedrock/tokens/count/response`	`api_key`, `body`, `model_id`
`ApplyGuardrail`	POST	`/guardrail/{value.json.guardrail_id}/version/{value.json.guardrail_version}/apply`	`ai/bedrock/guardrail/apply`	`ai/bedrock/guardrail/apply/response`	`api_key`, `body`, `guardrail_id`, `guardrail_version`

Deploy

Get the template card (JSON with template and metadata) from the broker, then deploy the connector:

mosquitto_pub -t '$SYS/Coreflux/Command' -m "-routeCode AMAZON_BEDROCK_RUNTIME"

The deployable connector from the REST API app, with BASE_ADDRESS already filled in:

DEFINE ROUTE AmazonBedrockRuntime WITH TYPE REST_API
    ADD REST_API_CONFIG
        WITH BASE_ADDRESS "https://bedrock-runtime.eu-west-1.amazonaws.com"
        WITH ENABLE_CLIENT "true"
        WITH USE_SSL "true"
        WITH REQUEST_TIMEOUT '120'
        WITH DEFAULT_CONTENT_TYPE "application/json"
    ADD EVENT Converse
        WITH SOURCE_TOPIC "ai/bedrock/converse/send"
        WITH DESTINATION_TOPIC "ai/bedrock/converse/send/response"
        WITH METHOD "POST"
        WITH ENDPOINT "/model/{value.json.model_id}/converse"
        WITH HEADERS
            ADD "Accept" "application/json"
            ADD "Content-Type" "application/json"
            ADD "Authorization" "Bearer {value.json.api_key}"
        WITH BODY "{value.json.body}"
    ADD EVENT InvokeModel
        WITH SOURCE_TOPIC "ai/bedrock/model/invoke"
        WITH DESTINATION_TOPIC "ai/bedrock/model/invoke/response"
        WITH METHOD "POST"
        WITH ENDPOINT "/model/{value.json.model_id}/invoke"
        WITH HEADERS
            ADD "Accept" "application/json"
            ADD "Content-Type" "application/json"
            ADD "Authorization" "Bearer {value.json.api_key}"
        WITH BODY "{value.json.body}"
    ADD EVENT CountTokens
        WITH SOURCE_TOPIC "ai/bedrock/tokens/count"
        WITH DESTINATION_TOPIC "ai/bedrock/tokens/count/response"
        WITH METHOD "POST"
        WITH ENDPOINT "/model/{value.json.model_id}/count-tokens"
        WITH HEADERS
            ADD "Accept" "application/json"
            ADD "Content-Type" "application/json"
            ADD "Authorization" "Bearer {value.json.api_key}"
        WITH BODY "{value.json.body}"
    ADD EVENT ApplyGuardrail
        WITH SOURCE_TOPIC "ai/bedrock/guardrail/apply"
        WITH DESTINATION_TOPIC "ai/bedrock/guardrail/apply/response"
        WITH METHOD "POST"
        WITH ENDPOINT "/guardrail/{value.json.guardrail_id}/version/{value.json.guardrail_version}/apply"
        WITH HEADERS
            ADD "Accept" "application/json"
            ADD "Content-Type" "application/json"
            ADD "Authorization" "Bearer {value.json.api_key}"
        WITH BODY "{value.json.body}"

Deploy it with -addRoute (do not include ADD METADATA).

Usage example

Call Converse by publishing a JSON payload to its request topic, and read the reply on the response topic:

mosquitto_sub -t 'ai/bedrock/converse/send/response' &
mosquitto_pub -t 'ai/bedrock/converse/send' -m '{"api_key": "<api_key>", "body": {}, "model_id": "<model_id>"}'

Quick Start

Coreflux MQTT Broker

LoT Language

Coreflux HUB

VS Code - LoT Notebooks

Coreflux Portal

Using AI with Coreflux

Product Updates

Overview

Auth

Template parameters

Events

Deploy

Usage example

​Overview

​Auth

​Template parameters

​Events

​Deploy

​Usage example

​Related

Overview

Auth

Template parameters

Events

Deploy

Usage example

Related