Amazon Bedrock AgentCore Long-Term Memory connector

With the Amazon Bedrock AgentCore Long-Term Memory outbound connector, you can retrieve persistent knowledge, such as facts, preferences, and summaries, from AWS Bedrock AgentCore Long-Term Memory in your BPMN process.

Prerequisites

To use the Amazon Bedrock AgentCore Long-Term Memory connector, you need the following:

• An AWS account with an access key and secret key, or a configured default credentials chain.
• An AgentCore Memory resource created in your AWS account.
• IAM permissions to execute the RetrieveMemoryRecords and ListMemoryRecords actions.

Learn more about Amazon Bedrock AgentCore Long-Term Memory in the official documentation.

note

Use Camunda secrets to store credentials and avoid exposing sensitive information directly from the process. Refer to managing secrets to learn more.

Create an Amazon Bedrock AgentCore Long-Term Memory connector task

You can apply a connector to a task or event via the append menu. For example:

• From the canvas: Select an element and click the Change element icon to change an existing element, or use the append feature to add a new element to the diagram.
• From the properties panel: Navigate to the Template section and click Select.
• From the side palette: Click the Create element icon.

After you have applied a connector to your element, follow the configuration steps or see using connectors to learn more.

Authentication

To authenticate, choose one of the methods from the Authentication dropdown. The supported options are:

• Credentials: Select this option if you have a valid pair of access and secret keys provided by your AWS account administrator.

note

This option is applicable for both SaaS and Self-Managed users.

• Default Credentials Chain: Select this option if your system is configured as an implicit authentication mechanism, such as role-based authentication, credentials supplied via environment variables, or files on target host. This approach uses the Default Credential Provider Chain to resolve required credentials.

note

This option is applicable only for Self-Managed or hybrid distributions.

Configuration

1. In the Region field, enter the AWS region where your AgentCore Memory resource is deployed. For example, us-east-1.
2. In the Memory ID field, enter the identifier of the AgentCore Memory resource you want to query.
3. In the Namespace field, enter a namespace prefix to scope memory records. For example, customer/12345. This is required by AWS to organize and isolate memory records.

Operations

The Amazon Bedrock AgentCore Long-Term Memory connector supports the following operations.

Retrieve memory records

Perform a semantic search to find relevant memory records based on a natural language query.

Parameters

Parameter	Required	Description
Search query	Yes	Semantic search query to find relevant memory records (up to 10,000 characters).
Memory strategy ID	No	Limits the search to memories created by a specific extraction strategy.
Max results	No	Maximum number of results to return (1–100). Defaults to 10.
Next token	No	Pagination token from a previous response to fetch the next page.

List memory records

List all memory records within the configured namespace.

Parameters

Parameter	Required	Description
Memory strategy ID	No	Filter memory records by a specific extraction strategy.
Max results	No	Maximum number of results to return (1–100). Defaults to 20.
Next token	No	Pagination token from a previous response to fetch the next page.

Response

The connector operations use the following response structure:

Field	Description
records	A list of memory record entries. Each entry contains the fields described below.
resultCount	The number of records returned.
nextToken	A token for retrieving additional results in subsequent requests. null if no more results.

Each record in the records list contains:

Field	Description
memoryRecordId	Unique identifier for the memory record.
content	The text content of the memory record.
memoryStrategyId	The extraction strategy that created this record.
namespaces	List of namespaces associated with the record.
createdAt	Timestamp when the record was created.
score	Relevance score (only present for retrieve operations).
metadata	Key-value metadata associated with the record.

Output mapping

1. Use Result Variable to store the response in a process variable. For example, memoryResult.
2. Use Result Expression to map specific fields from the response into process variables.

For example, to extract the content from all records:

= {
  memories: records.content,
  count: resultCount
}

Example response

The following is an example of the connector response:

{
  "records": [
    {
      "memoryRecordId": "mem-abc123",
      "content": "Customer prefers email communication over phone calls.",
      "memoryStrategyId": "preference-extraction",
      "namespaces": ["customer/12345"],
      "createdAt": "2024-01-15T10:30:00Z",
      "score": 0.95,
      "metadata": {
        "source": "support-ticket",
        "confidence": "high"
      }
    },
    {
      "memoryRecordId": "mem-def456",
      "content": "Customer is located in the Pacific timezone.",
      "memoryStrategyId": "fact-extraction",
      "namespaces": ["customer/12345"],
      "createdAt": "2024-01-10T14:20:00Z",
      "score": 0.87,
      "metadata": {
        "source": "profile-update"
      }
    }
  ],
  "resultCount": 2,
  "nextToken": null
}