generated: '2026-08-27'
method: searched
status: published
source: >-
  https://docs.confluent.io/cloud/current/ai/ai-tools/managed-mcp-server.html
  and https://github.com/confluentinc/mcp-confluent
description: >-
  Confluent ships TWO distinct MCP products. The managed servers are remote HTTPS endpoints an
  agent can POST to today, governed by the caller's existing Confluent Cloud RBAC. The
  open-source server is a local stdio process a human must install and run first, and Confluent
  states plainly that it is community-supported with no service level commitments. They are not
  interchangeable and are recorded separately below.
deployment:
  mode: both
  endpoint: https://api.confluent.cloud/mcp/v1
  install: npx -y @confluentinc/mcp-confluent
  package: https://www.npmjs.com/package/@confluentinc/mcp-confluent
  auth: api-key
  verified: probed
servers:
  - name: confluent-cloud-managed-global
    kind: managed
    transport: http
    url: https://api.confluent.cloud/mcp/v1
    auth: HTTP Basic with a Confluent Cloud API key (Cloud, global, or resource-scoped)
    scope: organization-wide operations not tied to a region
    probe:
      method: POST tools/list (jsonrpc 2.0)
      http_status: 401
      body: '{"errors":[{"status":"401","detail":"Unauthorized"}]}'
      note: >-
        The live tool schemas are auth-gated. The tool NAMES and descriptions below are
        transcribed verbatim from Confluent's published tool tables, not from an authenticated
        tools/list response, so inputSchema is not captured. Authenticated introspection is
        required to bind real input contracts.
  - name: confluent-cloud-managed-regional
    kind: managed
    transport: http
    url_template: https://mcp.<region>.<cloud>.confluent.cloud/mcp/v1/organizations/<org_id>
    auth: HTTP Basic with a resource-scoped Confluent Cloud API key
    scope: >-
      operations requiring direct access to resources in one region — topics and Schema
      Registry subjects. cloud is one of aws, gcp, azure.
    note: >-
      Templated by design; the concrete host depends on the customer's cloud, region and
      organization ID, so no single endpoint could be probed.
  - name: mcp-confluent
    kind: open-source
    transport: stdio
    install: npx -y @confluentinc/mcp-confluent
    package: https://www.npmjs.com/package/@confluentinc/mcp-confluent
    repo: https://github.com/confluentinc/mcp-confluent
    version: 1.5.0
    published: '2026-07-07'
    auth: api-key
    scope: >-
      Confluent Cloud, Confluent Platform and standalone Apache Kafka. Confluent's README
      states it provides "50+ tools across Kafka, Flink SQL, Schema Registry, Connectors,
      Tableflow, and more". Tool names are enumerable only by running
      `npx -y @confluentinc/mcp-confluent --list-tools`, which requires local execution, so
      they are not transcribed here rather than guessed.
    support: >-
      Community-supported. Confluent states it "doesn't provide dedicated support for it, and
      support is best-effort only, with no service level commitments."
tools:
  - name: list_environments
    server: managed-global
    category: discovery
    description: List all Confluent Cloud environments you can access, including environment IDs, names, and stream governance packages.
  - name: read_environment
    server: managed-global
    category: discovery
    description: Get details for a specific environment.
  - name: list_clusters
    server: managed-global
    category: discovery
    description: List Kafka clusters in an environment, including cluster type, cloud provider, region, and availability.
  - name: read_cluster
    server: managed-global
    category: discovery
    description: Get details and endpoints for a specific cluster.
  - name: list_connectors
    server: managed-global
    category: connectors
    description: List all connectors in a cluster with name, type, and current status.
  - name: get_connector_config
    server: managed-global
    category: connectors
    description: Get the full configuration for a connector, including configuration parameters and task list.
  - name: get_connector_status
    server: managed-global
    category: connectors
    description: Get runtime status and per-task status with error traces (PROVISIONING, RUNNING, DEGRADED, FAILED, PAUSED, DELETED).
  - name: get_connector_logs
    server: managed-global
    category: connectors
    description: Fetch recent connector logs. ERROR-level by default, optional WARN and INFO. Maximum lookback is 72 hours.
  - name: get_connector_offsets
    server: managed-global
    category: connectors
    description: Get partition and offset information showing the connector's consumption progress.
  - name: get_connector_metrics
    server: managed-global
    category: connectors
    description: Fetch time-series metrics for a connector.
  - name: get_connector_error_summary
    server: managed-global
    category: connectors
    description: Get an AI-generated summary of connector errors extracted from task status and error traces.
  - name: get_connector_fix_recommendations
    server: managed-global
    category: connectors
    description: Get AI-generated fix recommendations for a failing connector.
  - name: restart_connector
    server: managed-global
    category: connectors
    write: true
    description: Restart a connector and its tasks.
  - name: update_connector_config
    server: managed-global
    category: connectors
    write: true
    description: >-
      Update a connector's configuration. Pass only the fields you want to change; other fields
      are preserved. You can't remove a field by omitting it.
  - name: list_metrics
    server: managed-global
    category: metrics
    description: >-
      List available metric descriptors, including names, types, and units. Supports resource
      types kafka, connector, compute_pool, ksql, schema_registry.
  - name: query_metrics
    server: managed-global
    category: metrics
    description: >-
      Query time-series metric data for a cluster, such as throughput or consumer lag.
      Supports custom time intervals, granularity, and grouping by topic or consumer group.
  - name: list_kafka_topics
    server: managed-regional
    category: topics
    description: List all topics in a cluster.
  - name: describe_kafka_topic
    server: managed-regional
    category: topics
    description: Get topic configuration, including partition count and replication factor.
  - name: consume_kafka_messages
    server: managed-regional
    category: topics
    description: Read one to ten sample messages from a topic. You can target a specific partition and offset.
  - name: list_schema_subjects
    server: managed-regional
    category: schema-registry
    description: List all schema subjects in Schema Registry, including subject name, version, schema ID, and schema type.
  - name: read_schema_subject
    server: managed-regional
    category: schema-registry
    description: Get the full schema definition for a subject, including content, version, and references.
safety:
  read_only_by_default: true
  note: >-
    Confluent states most managed MCP tools are read-only and that you cannot create, change or
    delete clusters or topics through MCP tools. The two exceptions are restart_connector and
    update_connector_config, which change Confluent Cloud resources.
  authorization: >-
    MCP tools return only the resources the caller's API key has permission to access; access is
    governed by existing Confluent Cloud RBAC.
  audit: >-
    MCP tool calls appear in the Confluent Cloud audit log, logged with a method name prefixed
    `mcp.tools/call.` followed by the tool name.