--- title: "List Organization Agent Memory Servers" method: GET path: "/v4/organizations/{organizationId}/aiServices/agentMemoryServers" tags: ["AI Agent Memory Servers"] --- # List Organization Agent Memory Servers `GET /v4/organizations/{organizationId}/aiServices/agentMemoryServers` ## Request - Base URL: `https://cloudapi.cloud.couchbase.com` - URL: `https://cloudapi.cloud.couchbase.com/v4/organizations/{organizationId}/aiServices/agentMemoryServers` - Auth: HTTP bearer - TypeScript: (a function with typed inputs; its response is checked against the schema) Lists every agent memory server in the organization that the caller has access to. Use the `projectId` query parameter to narrow the results to a single project, and the `search` query parameter to match against the name and description. In order to access this endpoint, the provided API key must have at least one of the following roles: - Organization Owner - Project Creator - Project Owner - Project Manager - Project Viewer - Database Data Reader/Writer - Database Data Reader - Cluster Owner - Cluster Data Writer - Cluster Data Viewer To learn more, see [Organization, Project, and Database Access Overview](https://docs.couchbase.com/cloud/organizations/organization-projects-overview.html). ## Query parameters - `projectId` string, uuid - `search` string - `page` integer - `perPage` integer - `sortBy` string[] - `sortDirection` 'asc' | 'desc' - `status` 'pending' | 'deploying' | 'healthy' | 'unhealthy' | 'deploymentFailed' | 'destroying' | 'destroyFailed' ## Response `200` Successfully listed the agent memory servers for the organization. - ListAgentMemoryServersResponse - `data` GetAgentMemoryServerResponse[], required: The agent memory servers found by the given query. Empty if the query matched no agent memory server. - `id` string, uuid, required: The GUID4 ID of the agent memory server. - `name` string, required: The name of the agent memory server. - `description` string: A short description of the agent memory server. - `tenantId` string, uuid, required: The GUID4 ID of the organization the agent memory server belongs to. - `projectId` string, uuid, required: The GUID4 ID of the project the agent memory server belongs to. - `clusterId` string, uuid, required: The GUID4 ID of the cluster the agent memory server is deployed against. - `config` GetAgentMemoryServerConfig, required: The deployment and processing configuration of the agent memory server. - `nodeConfig` AgentMemoryServerNodeConfigResponse, required - `instanceType` 'small' | 'medium' | 'large', required: The T-shirt size of the compute instances backing the agent memory server. Only `small` is offered for new agent memory servers; an existing server reports the size it was deployed with, which may be `medium` or `large`. - `nodeCount` integer, required: The number of nodes deployed for the agent memory server. - `provider` string, required: The cloud service provider the agent memory server is deployed on. Inherited from the cluster. - `region` string, required: The cloud provider region the agent memory server is deployed in. Inherited from the cluster. - `bucket` string, required: The name of the bucket on the cluster in which the agent memory is stored. This is the bucket name, not the URL-safe base64 `bucketId` used in v4 paths. - `memoryBlockTTL` integer, nullable: The time to live, in seconds, applied to each memory block. Returns `null` when memory blocks never expire. - `embedding` AgentMemoryServerEmbeddingConfig, required: The embedding model used to vectorize memory blocks. - `integrationId` string, required: The ID of the AI provider integration that serves the embedding model. The ID comes from the List providers endpoint, `GET /v4/organizations/{organizationId}/aiServices/providers`. - `model` string, required: The name of the embedding model, as offered by the referenced integration. - `llm` AgentMemoryServerLLMConfig, required: The large language model used to summarize and extract memory. - `integrationId` string, required: The ID of the AI provider integration that serves the language model. The ID comes from the List providers endpoint, `GET /v4/organizations/{organizationId}/aiServices/providers`. - `model` string, required: The name of the language model, as offered by the referenced integration. - `semanticExtraction` AgentMemoryServerSemanticExtraction: Controls semantic extraction of structured facts from stored memory. - `enabled` boolean: Whether semantic extraction is enabled for the agent memory server. - `advancedConfig` AgentMemoryServerAdvancedConfig: Advanced tuning knobs for the agent memory server's processing pipeline. Every field is optional; omitted fields fall back to the documented default. - `summaryTokenLimit` integer: The maximum number of tokens the language model may emit when summarizing a memory block. Values below the minimum truncate the model's response before it finishes emitting the summary, which fails the extraction for that block. - `processingBatchSize` integer: The number of memory blocks processed together in a single batch. - `maxRequestsPerMinute` integer: The maximum number of requests the agent memory server issues to the model providers per minute. The default is a conservative starting point rather than a recommendation. Set it to the rate limit of your own provider tier for the configured model, since that limit, and not this field, is the real ceiling. In practice `maxTokensPerMinute` is the tighter of the two limits for most workloads. On a multi-node agent memory server this is the total across the deployment, divided across the live nodes. - `maxTokensPerMinute` integer: The maximum number of tokens the agent memory server sends to the model providers per minute. The default is a conservative starting point rather than a recommendation. Set it to the token limit of your own provider tier for the configured model, since that limit, and not this field, is the real ceiling. On a multi-node agent memory server this is the total across the deployment, divided across the live nodes. Each node therefore receives only `maxTokensPerMinute / nodeCount`, and a request larger than a node's share can never be dispatched no matter how long it waits. The configuration must satisfy: maxTokensPerMinute >= perRequestTokenLimit * nodeCount Requests that violate this are rejected at creation time. - `maxConcurrentTasks` integer: The maximum number of processing tasks that run concurrently. This is a safety backstop rather than a throughput setting. The rate limits above are usually reached long before this one is. - `perRequestTokenLimit` integer: The maximum number of tokens allowed in a single model request. This value is bounded by the per-node share of the token budget, so it must satisfy: perRequestTokenLimit <= maxTokensPerMinute / nodeCount - `idleSleepSeconds` number, double: The number of seconds the agent memory server waits between dispatch cycles. - `queueMaxSize` integer: The maximum number of items allowed in the processing queue. - `maxFailCount` integer: The number of times a task is attempted before it is marked as failed. Set it to `1` to fail a task on its first error without retrying. - `taskTimeoutSeconds` integer: The number of seconds after which an in-flight processing task times out. - `network` AgentMemoryServerNetwork, required: The network details used to reach the agent memory server. - `endpoint` string: The HTTPS endpoint that agent applications use to reach the agent memory server. - `status` 'pending' | 'deploying' | 'healthy' | 'unhealthy' | 'deploymentFailed' | 'destroying' | 'destroyFailed', required: The current deployment status of the agent memory server: * `pending`: The agent memory server record has been created but deployment has not started. * `deploying`: The agent memory server is being deployed. * `healthy`: The agent memory server is deployed and serving normally. * `unhealthy`: The agent memory server is deployed but is failing its health checks. * `deploymentFailed`: The deployment of the agent memory server failed. * `destroying`: The agent memory server is being destroyed. * `destroyFailed`: The destruction of the agent memory server failed. - `audit` CouchbaseAuditData, required - `createdBy` string, required: The user who created the resource; this will be a UUID4 ID for standard users and will be a string such as "internal-support" for internal Couchbase support users. - `createdAt` string, date-time, required: The RFC3339 timestamp associated with when the resource was initially created. - `modifiedBy` string, required: The user who last modified the resource; this will be a UUID4 ID for standard users and will be a string such as "internal-support" for internal Couchbase support users. - `modifiedAt` string, date-time, required: The RFC3339 timestamp associated with when the resource was last modified. - `version` integer, required: The version of the document. This value is incremented each time the resource is modified. - `cursor` Cursor, required - `pages` Pages, required - `page` integer, required: Current page of results, starting from page 1. - `next` integer: Number of the next page of results. Not set on the last page. - `previous` integer: Number of the previous page of results. Not set on the first page. - `last` integer, required: Number of the last page of results. - `perPage` integer, required: Number of items displayed in each page. - `totalItems` integer, required: Total items found by the given query. - `hrefs` Hrefs, required - `first` string, uri, required: The base URL, endpoint, and path parameters required to fetch the first page of results. - `last` string, uri, required: The base URL, endpoint, and path parameters required to fetch the last page of results. - `previous` string, uri, required: The base URL, endpoint, and path parameters required to fetch the previous page of results. Empty if there is no previous page. - `next` string, uri, required: The base URL, endpoint, and path parameters required to fetch the next page of results. Empty if there is no next page. ## Other responses - `400`: Returned when we are unable to decode the recevied payload. - `401`: The client does not have the valid credentials to access this resource. - `403`: The client does not have the necessary permissions to access this resource. - `404`: The requested resource was not found. - `429`: Returned when the client exceeds the rate limit for the given APIKey. - `500`: An unexpected error occurred in the server while processing this request. ## Changes - **2026-10-19** `a5e88bcf71cb`: 1 info - endpoint added [Operation history](https://skmtc.dev/couchbase/apis/couchbase-capella-management-api/changes/v4/organizations/:organizationId/aiServices/agentMemoryServers/get.md) --- [API](https://skmtc.dev/couchbase/apis/couchbase-capella-management-api.md) - [All operations](https://skmtc.dev/couchbase/apis/couchbase-capella-management-api/llms.txt) - [OpenAPI document](https://skmtc.dev/couchbase/apis/couchbase-capella-management-api/revisions/a5e88bcf71cb?raw)