Skip to main content
To discover information about your Kafka resources, you can connect to Conduktor Console from an AI assistant via the Model Context Protocol (MCP). Use Console UI for setup instructions: Setup MCP
The MCP server uses your own permissions based on the token you generate. Make sure the token has the appropriate access rights for the resources you want to query.
Use AI tools for discovery and exploration only — results are open to interpretation and outside of Conduktor’s control.

​
MCP tools

All tools are read-only. They cover Kafka resources and Insights data held by Console.

​
Cluster management

  • list-clusters: list all Kafka clusters available in Conduktor
  • get-cluster: get details about a specific cluster by its slug

​
Topic operations

  • list-topics-with-usage: list topics with usage metrics (message count, size, rates) and filtering by name, cleanup policy, labels, tags, or to hide internal/stream topics
  • get-last-messages: retrieve the last N messages from a topic (up to 100 messages, all partitions or a specific subset)

​
Insights

  • insights-cluster: cluster health summary and serialization format breakdown
  • insights-topics: full topic insights — partition skew, replication issues, partition distribution, health overview, empty/stale/tiny topics, schema coverage, and Self-service coverage

​
Schema registry

  • list-subject-names: list subject names in a cluster, with filtering by schema type (AVRO, PROTOBUF, JSON)

​
Consumer groups

  • list-consumer-groups: list consumer groups with state, lag, and member count; filter by name, state, or labels; sort by lag, members, etc.
  • list-consumer-groups-by-topic: list consumer groups consuming from a specific topic

​
Gateway and Interceptors

  • list-interceptors: list interceptors configured in a cluster (requires Conduktor Gateway)