Remotion Studio lets AI agents control video editing

Remotion has updated its browser-based React video editing interface, Remotion Studio, to support WebMCP starting in version 4.0.518. This integration allows AI agents to interact directly with the video editor using structured, typed functions instead of relying on brittle computer-use methods like taking screenshots and guessing where to click. WebMCP is a W3C incubation draft co-developed by Google's Chrome team, Microsoft's Edge team, and Alex Nahas, with native browser support anticipated by mid-to-late 2026.

Under this new system, an agent can read and write to the editor without needing an external server. Read tools allow agents to inspect compositions, playback states, and element outlines. For example, the get_canvas_html tool retrieves the composition HTML up to a 100,000-character limit, while get_selection identifies the source file and line of a selected timeline item. Write tools enable agents to play, pause, mute, and manage guides. The set_playback_rate tool restricts speed adjustments to multipliers between -4 and 4, and seek_to_frame automatically clamps out-of-range values to the final frame.

For developers and video editors, this shifts the paradigm of AI assistance from visual guessing to deterministic API control. Instead of hoping an agent clicks the right pixel, users can prompt a tool to perform precise actions like adding a vertical guide at 640 pixels. However, adoption is currently limited. Among major agent frameworks, only ChatGPT Codex currently implements WebMCP, meaning the integration remains dormant for users on other platforms. Additionally, while Remotion limits its write tools to timeline navigation to prevent unauthorized filesystem writes, the WebMCP specification still warns of security threats like prompt injection and privacy leakage.