Enables AI agents to directly manipulate CapCut project files via the Model Context Protocol.
Advertisement
CapCut provides a local Model Context Protocol (MCP) server, built with Elixir and OTP, that allows AI agents like Claude to directly read and edit CapCut's local JSON project files. It bypasses the need for a formal CapCut API by interacting directly with the project file structure, offering tools for managing video drafts, adding text, and incorporating media clips. This enables advanced AI-driven automation for video editing workflows, all within a robust and fault-tolerant Elixir environment.
Key Features
01Built with Elixir/OTP for concurrency and fault-tolerance
020 GitHub stars
03Support for removing clips by segment ID
04Direct manipulation of CapCut project JSON files
05Ability to add text overlays and media clips to timelines
06Tools for listing, inspecting, creating, and editing projects
Use Cases
01Automated creation and modification of CapCut video drafts by AI agents
02Integrating AI into video editing workflows for dynamic content generation
03Programmatic management of local CapCut projects
