Confluent MCP Server
Confluent MCP server for Kafka: topics, messages, Flink SQL, Schema Registry and connectors
Install
npx @confluentinc/mcp-confluent --init-config
npx @confluentinc/mcp-confluent --config ./config.yamlFirst create config.yaml and fill in only the service blocks you need.
This is third-party code. Review the repository files before installing.
What it does
The server connects an agent to Confluent Cloud, Confluent Platform or plain Apache Kafka. Against local Kafka it handles topics, producing and consuming messages, consumer groups and schemas. Confluent Cloud adds Flink SQL with diagnostics, Kafka Connect connectors, catalog tags, Tableflow, metrics and billing. Tools are enabled by blocks in config.yaml, and diagnostic tools explain why a tool is missing.
Who it is for. For data engineers and developers working with Kafka and Confluent.
Good fit when
- You need to inspect topics, consumer lag or read messages
- You need to write and debug a Flink SQL statement
- You need to find out why a connector failed
Not a fit when
- The cluster is production and the agent must not delete topics or connectors
- You already use Confluent Cloud's managed MCP server and need no extensions
Example request
Show consumer lag for the orders-service group across topics and read the last five messages from ordersLimitations
Requires Node.js 22.19 or newer and access to Kafka or a Confluent Cloud account with keys or OAuth login. It is community-supported: Confluent gives no guarantees. Some Cloud tools still work only with static keys.
How to disable. Remove the server from the client's MCP config. Revoke Confluent Cloud keys in the console.
MCP
- Transport
- stdio
- Authentication
- API key
Security check
- Can delete topics, schemas and connectors
- Produces messages to clusters, including production
README in short
The README starts with an npx quick start: generate config.yaml and run the server. Tools are grouped by availability: always on, Confluent Cloud and local deployments. It covers Confluent Platform, TLS with internal CAs, OAuth login and client setup. MIT licensed with community support.
FAQ
Does it work without Confluent Cloud?
Yes, with local Kafka and Schema Registry you get topic, message, group and schema tools.
How do I see which tools are enabled?
Run npx -y @confluentinc/mcp-confluent --list-tools or use the explain-disabled-tools tool.
Related
166 skills for scientific work: bioinformatics, cheminformatics, clinical data, geospatial analysis and 100+ databases
Google's open-source MCP server for databases: ready tools for Postgres, MySQL, BigQuery, Spanner and more, plus custom tools in tools.yaml
Official Hugging Face skills: Hub operations via the hf CLI, datasets, model training, Spaces, evals and deployment
A token-lean MCP server for PostgreSQL, MySQL, MariaDB, SQL Server and SQLite with a read-only mode