mcp-kafka mcp

v1.0.1 · MCP Tool · messaging · registry.pascalai.org

Apache Kafka via Confluent REST Proxy API (HTTP-based, no driver required). Produces and consumes messages, manages topics.

kafkamessagingstreamingeventsconfluent

Input Parameters

Parameter	Type	Description
operationrequired	string	Operation to perform. One of: `list_topics`, `get_topic`, `create_topic`, `delete_topic`, `produce`, `consume`, `list_consumers`, `get_offsets`.
baseUrlrequired	string	Confluent REST Proxy base URL (e.g. http://localhost:8082).
apiKeyoptional	string	API key for Confluent Cloud (used as Basic auth username).
apiSecretoptional	string	API secret for Confluent Cloud (used as Basic auth password).
clusterIdoptional	string	Kafka cluster ID for v3 API operations (create_topic, delete_topic). e.g. lkc-xxxxx.
topicoptional	string	Topic name. Required for: get_topic, create_topic, delete_topic, produce, consume, get_offsets.
groupoptional	string	Consumer group name. Required for: list_consumers.
instanceoptional	string	Consumer instance name within the group.
partitionoptional	integer	Partition number (default: 0). Default: `0`.
offsetoptional	integer	Start offset (default: 0). Default: `0`.
countoptional	integer	Number of messages to fetch (default: 10). Default: `10`.
bodyoptional	string	JSON body string for produce or create_topic.
keyoptional	string	Message key string for produce.

Output Fields

Field	Type	Description
ok	boolean
data	object
deleted	string
error	string

Examples

List all topics

// Input
{
  "operation": "list_topics",
  "baseUrl": "http://localhost:8082"
}

// Output
{
  "ok": true
}

Produce a message

// Input
{
  "operation": "produce",
  "baseUrl": "http://localhost:8082",
  "topic": "orders",
  "body": "{\"records\":[{\"value\":\"hello\"}]}"
}

// Output
{
  "ok": true
}

Install & Discovery

Install

ppm install mcp-kafka

Get JSON Schema

GET /v1/packages/mcp-kafka/1.0.1/schema

Discover by keyword

GET /v1/mcp/discover?q=kafka

Discovery hint: Install with ppm install mcp-kafka or invoke remotely via POST /v1/invoke/mcp-kafka on the MCP Service.

PascalAI Usage

uses toolslib;
var Tool := LoadTool('mcp-kafka');
var R := Tool.Call(JsonObj(['operation','list_topics','baseUrl','http://localhost:8082']));
Writeln(R.ToJSON);