An MCP (Model Context Protocol) server for ZapCap that exposes tools to upload videos, create processing tasks, and monitor progress via the ZapCap API. This unofficial implementation enables token-based authentication via a single environment variable (ZAPCAP_API_KEY) and can be run locally or in Docker. Requirements include uv and a ZapCap API key. Install via an MCP client by adding the server to your mcp.json, or install the uv-based tool with 'uv tool install zapcap-mcp-server'. You can run the server with a Docker image (bogdan01m/zapcap-mcp-server:latest) and pass ZAPCAP_API_KEY as an environment variable. The server provides tools: zapcap_mcp_upload_video, zapcap_mcp_upload_video_by_url, zapcap_mcp_get_templates, zapcap_mcp_create_task (highly configurable: auto_approve, language, enable_broll, broll_percent, subtitle options, style options), and zapcap_mcp_monitor_task. It supports a natural-language interface to describe desired actions and validates parameters with Pydantic for type safety. This design enables simple, language-driven automation for ZapCap’s video captioning and B-roll generation workflows, with token-management benefits over direct API usage.