Skip to main content
The Remote StreamNative MCP Server provides a hosted Model Context Protocol (MCP) service at https://mcp.streamnative.cloud. To connect, you must construct a cluster-specific endpoint by appending /mcp/x/<organization>/<pulsar-instance>/<cluster-id> to the base URL. It is built on the open-source StreamNative MCP Server and offered as a preview service so you can consume managed tools without running infrastructure.

​
Instant access from your IDE

  • Launch tooling with a single click in supported editors such as VS Code—no local binaries or runtimes required.
  • Construct your cluster endpoint and paste it into any MCP-compatible client to reuse the same catalog across teams and environments.
  • Share connections through workspace templates so new collaborators adopt a consistent tool set.

​
Why use the remote server

  • Skip local installation and upgrades—the managed endpoint stays current with the latest StreamNative MCP capabilities.
  • Connect from IDEs such as VS Code with a single click or point any MCP-compatible runtime to your cluster endpoint.
  • Authenticate with the same StreamNative Cloud credentials you already use for agent deployments.

​
StreamNative Cloud integration

The managed server authenticates against StreamNative Cloud and understands your organization context. After you sign in, you can:
  • Browse and switch between clusters that belong to your organization.
  • Inspect the status of cluster resources through natural language prompts.
  • Interact with StreamNative services using your existing service accounts and permission model.

​
Platform support

The Remote StreamNative MCP Server exposes tools for both Apache Kafka and Apache Pulsar workloads.
  • Apache Kafka — manage topics, partitions, consumer groups, schema registry artifacts, Kafka Connect deployments (preview), and run client operations such as producing and consuming messages.
  • Apache Pulsar — administer tenants, namespaces, topics, and schemas; run client operations; and manage Functions, Sources, and Sinks in read-only mode.
  • Orca integrations — surface Pulsar Functions and Orca agents as MCP tools so downstream agents can invoke them directly.

​
Always current and managed

  • StreamNative operates the service so you automatically receive server updates, new tools, and security patches.
  • The remote environment mirrors the latest StreamNative Cloud features, keeping documentation and runtime behavior aligned.
  • Preview status means feedback can shape the roadmap—share requests through support channels to influence future releases.

​
Security and authorization

  • OAuth 2.1 authorization flows align with the evolving MCP authorization specification, allowing interactive sign-in from IDEs and browsers.
  • StreamNative Cloud service account API keys offer non-interactive access for automated clients.
  • The server is multi-tenant aware and scopes tool discovery to the resources your identity can access.

​
Getting started

  1. Build the cluster endpoint by appending /mcp/x/<organization>/<pulsar-instance>/<cluster-id> to the base URL. For example, organization o-sndev, instance instance-1, cluster c-cluster resolves to https://mcp.streamnative.cloud/mcp/x/o-sndev/instance-1/c-cluster.
  2. When using VS Code, paste the cluster endpoint URL into your MCP client configuration and complete the OAuth 2.1 sign-in or supply a service account API key.
  3. For other MCP-compatible runtimes, configure the cluster endpoint in the same way you would add any MCP server and complete the authentication step.
Once connected, your IDE or agent runtime receives Server-Sent Events (SSE) notifications whenever StreamNative adds or updates tools, keeping the catalog in sync without manual refreshes.