Bitmovin’s Stream Lab now includes an MCP Server that gives AI agents and large language models (LLMs) the ability to perform video playback tests and analyse results in real-time.
Bitmovin’s Stream Lab now includes an MCP Server that gives AI agents and large language models (LLMs) the ability to perform video playback tests and analyse results in real-time. Stream Lab is an automated tool for video streaming services, allowing them to use Quality of Experience (QoE) analytics to test users’ devices. It gives insights into key metrics such as startup time, bitrate, buffering and error incidents, helping users understand viewer experiences and identify the necessary optimisations for improving stream quality.
The Model Context Protocol (MCP) open standard defines a structured interface between AI agents and external systems. Unlike a traditional API, which can only work through fixed endpoints, an MCP-based server acts as a purpose-built orchestration layer for AI interactions.
Bitmovin designed their new server together with Alpic, a developer of open-source, cloud-based tools for building and scaling MCP apps. The Stream Lab MCP Server allows AI agents and LLMs to use Stream Lab without integration code or manual test setup. Using natural language commands, AI agents can carry out practical tasks such as creating and managing test setups, running tests and analysing playback on over 30 physical test device targets including Samsung, LG and Vizio TVs, web browsers and games consoles.
The Stream Lab MCP Server can also be combined with Bitmovin's Observability MCP Server to give AI agents access to use real-world QoE data to enhance testing with Stream Lab. This ability could be used to prioritise tests on devices where viewers are seeing the most issues and measuring the impact of changes, for example. By using simple questions as prompts, the result would be a natural-language exploration of video playback metrics, license usage and viewer experience insights.
The MCP server connects Bitmovin accounts to an AI assistant. Once configured, the assistant could be questioned on, for instance, average video startup times in particular locations, suggestions for improvement based on earlier data sets, document summaries or average bitrates within a specified period. These questions are translated into precise API calls, returning structured analytics data that the assistant can interpret or visualise.
James Varndell, Senior Director of Product Management, Bitmovin, commented, "Bitmovin's automated video stream testing tool Stream Lab already replaces manual testing processes with fully automated testing on physical devices. Now, with the launch of the Stream Lab MCP, users have an effective new way to control playback testing using simple commands in everyday language without human intervention.
“With Stream Lab's MCP Server, video services can enable MCP-compatible AI tools such as Claude to interact with Stream Lab tests using natural language to explore test devices, create and run tests, and analyze test reports. This makes agentic AI practical at scale." bitmovin.com
