v1

latestOpenAPI 3.0.22026-07-173834119.9 KB
Model Invocation

Async Invoke Model Endpoint

Endpoint to Invoke a Model on Bedrock Asynchronously

This endpoint allows users to invoke a model on Bedrock asynchronously using either a simple text prompt or a series of messages. It returns an invocation ID that can be used to retrieve the result later.


Request Body

ParameterTypeDescription
model_namestrThe name of the model to invoke. Must be one of the supported models.
promptstrA simple text prompt (str) or a list of messages (see below for message format).
max_tokensOptional[int]The maximum number of tokens to generate in the response.
temperatureOptional[float]Sampling temperature to use. Higher values make the output more random.
top_pOptional[float]Probability threshold for nucleus sampling.
top_kOptional[int]The number of highest probability vocabulary tokens to keep for top-k filtering.
stop_sequencesOptional[List[str]]Sequences where the generation will stop.

Example:

{
    "model_name": "example_model",
    "prompt": "Translate the following text to French: 'Hello, how are you?'",
    "max_tokens": 100,
    "temperature": 0.7,
    "top_p": 0.9,
    "top_k": 50,
    "stop_sequences": ["\n"]
}

Response Body

FieldTypeDescription
invocation_idstrThe ID of the asynchronous invocation.

Errors

  • 400 Bad Request: If the request parameters are invalid.
  • 500 Internal Server Error: If there is an unexpected error during model invocation.
post/model/async_invoke

Request body

model_namestring required
max_tokensinteger
temperaturenumber
top_pnumber
top_kinteger
stop_sequencesstring[]
system_promptsobject[]

Example request

{
  "system_prompts": [
    {
      "text": "You are a helpful assistant."
    }
  ]
}

Response

Successful Response

{"stackTrail":"paths:/model/async_invoke:post:responses:200:content:application/json:schema","oasType":"schema","type":"unknown"}
All 38 operations