The Wowza Streaming Engine MCP Server provides a Model Context Protocol (MCP) interface for interacting with the Wowza Streaming Engine via natural-language or structured API calls.
It acts as a middleware service between AI or automation clients (agents) and the Wowza Streaming Engine REST API, enabling programmatic control over streaming operations such as starting, stopping, and configuring live applications.
Note: The Wowza Streaming Engine MCP Server is currently in beta. It provides early access to the MCP server and allows developers to explore AI-powered control of Wowza Streaming Engine through the Model Context Protocol (MCP). As a beta product, it's still under active development. Some features or configuration options may be limited. Additionally, there may be changes to the API and deployment process in future releases. Your feedback during this phase helps improve stability, performance, and usability.
Overview
The Wowza Streaming Engine MCP Server is a Python-based companion service built with FastAPI that implements the Model Context Protocol (MCP).
It translates MCP requests into Wowza Streaming Engine REST API calls. This allows tools and assistants that support MCP to send high-level intent commands — for example, “start the live stream” — which the server executes using the underlying Wowza Streaming Engine REST API.
Use cases
The following use cases show what you can do when integrating the Wowza Streaming Engine MCP Server with your AI assistants.
| Category | Description |
|---|---|
| Live stream management | Create, retrieve, update, and delete live stream sources and publishers. |
| nDVR operations | List, inspect, and convert nDVR stores to MP4 files for playback or archiving. |
| Stream targets (push publishing) | Create, list, and remove stream targets to distribute live content across multiple destinations. |
| Media Cache management | Configure cache sources and manage cached assets using the Media Cache v2 and v3 APIs. |
| Analytics and statistics | Retrieve current and historical performance metrics for applications and the server. |
| File management | Create, edit, and manage SMIL and stream files used in live or on-demand workflows. |
| Recording control | Start, list, split, and stop stream recordings for content capture and replay. |
| Server administration | Manage user accounts, server access, and log retrieval operations. |
| Application management | Create and configure Wowza Streaming Engine applications and update their settings. |
| Transcoder operations | Manage transcoder templates, configurations, and active transcoding sessions. |
| Server monitoring | Retrieve server health data, performance tuning, and monitoring configurations. |
| MediaCaster management | Configure and manage MediaCaster connections for live stream ingest. |
| Listener configuration | Adjust listener and network interface settings for optimized communication. |
| License management | View, validate, and manage Wowza Streaming Engine licenses. |
| Logging configuration (Log4j) | Configure Log4j settings to control server log output and verbosity. |
| VHost management | Configure and maintain virtual hosts for isolated or multi-tenant streaming environments. |
Core components
The MCP server comprises several layers that work together to translate MCP requests into Wowza Streaming Engine REST API calls.
- MCP API layer – A FastAPI-based service that defines the MCP schema and endpoints, handling inbound requests from compatible clients. An authentication layer within this API layer securely passes credentials when the MCP server interacts with Wowza Streaming Engine. When the client initiates a request, the MCP server injects the appropriate authentication headers — such as configured username and password — into the REST API call. This enables Wowza Streaming Engine to validate the request without exposing sensitive credentials to the client or requiring separate login steps.
- MCP adapter layer – The adapter layer translates MCP commands into Wowza Streaming Engine REST API calls, ensuring consistent, structured communication. It defines how tools—such as API endpoints, operations, and management commands—are exposed to large language models (LLMs) through the Model Context Protocol. Behind the scenes, it maps each operation (for example, creating a live stream or managing applications) to an MCP-compatible schema that defines inputs, outputs, and execution logic. When an LLM sends a request, the adapter layer converts it into a REST call, executes it, and returns a response that the model can interpret.
Deployment
The MCP server is distributed as a Docker image that supports MCP's STDIO and streamable-HTTP protocols. You can run the Docker container alongside Wowza Streaming Engine within the same Docker Compose network.
Once deployed, the MCP container automatically connects to Wowza Streaming Engine’s REST API endpoint, inheriting authentication and configuration settings from the Docker Compose file. Running both services together ensures synchronized startup, shared network isolation, and unified management within a single deployment environment.
For more information and to get started, see:
- Get started with the Wowza Streaming Engine MCP Server
- Wowza Streaming Engine MCP Server configuration reference
Notes:
- We recommend using the Docker Compose deployment for the Wowza Streaming Engine MCP server. This setup includes preconfigured Docker images where access to the Wowza Streaming Engine REST API is already enabled, Swagger ports are configured, and REST API authentication is set using the basic authentication method. This approach simplifies setup and ensures a consistent, ready-to-use environment for MCP server integration.
- The Docker images used in the Docker Compose deployment — available at wowza/wowza-streaming-engine-manager and wowza/wowza-streaming-engine — are only compatible with Wowza Streaming Engine version 4.9.3 and later. Earlier versions are not supported.
Integrations with external clients
This documentation focuses on connecting the Wowza Streaming Engine MCP Server with Visual Studio Code (VS Code), allowing you to use natural-language or structured commands to manage Wowza Streaming Engine operations such as starting applications, managing live streams, or viewing server statistics—all without leaving your editor. For more, see:
You can also connect to the MCP server by using other MCP-compatible clients, such as Windsurf and Cursor. These tools can communicate with the Wowza Streaming Engine MCP Server through the same standardized protocol, enabling flexible integration across different workflows and environments. For more, see this README.
Finally, we also provide an integration that leverages the Google Agent Development Kit (ADK) Client to build or extend custom agents that can connect to the Wowza Streaming Engine MCP Server.
