Invoke Model

Invoke an AWS Bedrock foundation model to generate text, embeddings, or other model outputs

Use it in a workflow

Add a step and choose the AWS Bedrock connector.
Pick the Invoke Model action (under bedrock).
Fill in the fields below, then reference the result from later steps as {{invokeModel.response.data}}.

Fields

modelIdstringrequired

The model identifier or ARN. Can be a Model ID (e.g., anthropic.claude-haiku-4-5-20251001-v1:0), a Model ARN, or an Inference Profile ARN — check your AWS Bedrock console's model catalog for what's enabled and current in your account/region.

contentTypestring

The MIME type of the input data

Default "application/json"

acceptstring

The desired MIME type of the output

Default "application/json"

bodystringrequired

JSON string containing the model input. Format depends on the model (e.g., Anthropic uses messages array, Titan uses inputText)

maxTokensnumber

Maximum number of tokens to generate (optional, may be included in body)

temperaturenumber

Sampling temperature (0.0 to 1.0). Higher values make output more random (optional, may be included in body)

topPnumber

Nucleus sampling parameter (0.0 to 1.0). Controls diversity (optional, may be included in body)

topKnumber

Top-K sampling parameter. Number of highest probability tokens to consider (optional, may be included in body)

stopSequencesstring

JSON array of strings that stop generation when encountered (e.g., ["\n\nHuman:", "\n\nAssistant:"])

Need more? See the AWS Bedrock guide for connection setup and behaviour shared by every action.

Was this page helpful?