creatify-mcp is an enterprise-grade MCP (Model Context Protocol) server for Creatify AI, designed to provide the capabilities of the Creatify AI video generation platform to AI assistants, chatbots, and automation tools in the MCP ecosystem. Built on top of the @tsavo/creatify-api-ts TypeScript client library, it turns complex video creation workflows into simple natural language interaction.
-
Main purpose Through this MCP server, AI assistants can understand and execute video generation tasks. For example, a user can say to Claude Desktop: "Create a 16:9 Anna avatar video with the line 'Welcome to our product demo' and wait for completion," and the server can actually execute it. It makes AI video generation capabilities easier for developers and AI assistants to use.
-
Main capabilities
-
12 MCP tools: create_avatar_video (create an AI avatar video with lip sync), create_url_to_video (convert a website into a professional video), generate_text_to_speech (generate natural speech from text), create_multi_avatar_conversation (create a video of multiple avatars conversing), create_custom_template_video (generate a video using a custom template), create_ai_edited_video (automatically edit and enhance videos), create_ai_shorts (create short videos suitable for TikTok and Instagram Reels), generate_ai_script (generate an AI script for a video), create_custom_avatar (design and create a custom avatar, i.e. DYOA), manage_music (upload, manage, and use background music), create_advanced_lipsync (advanced lip sync with emotion and gesture control), how_to_use (get detailed usage information for any tool), get_video_status (check the status of a video generation task).
-
6 MCP resources: creatify://avatars (list of available AI avatars), creatify://voices (list of available voices), creatify://templates (available custom video templates), creatify://music (available background music library), creatify://credits (remaining API credits), creatify://avatar/{avatarId} (detailed information about a specific avatar).
-
5 workflow prompt templates: create-product-demo (product demo), create-social-content (social media content), create-educational-video (educational tutorial video), create-marketing-campaign (marketing promotional video), analyze-video-performance (video analysis and optimization).
-
Advanced MCP features: reusable video creation templates and workflows (Prompts), structured logging with multiple severity levels (debug, info, notice, warning, error, critical, alert, emergency), real-time progress tracking during video generation, a how_to_use self-describing help tool for AI assistants, real-time status updates and progress notifications.
III. How to Use Prerequisites: Node.js 18 or higher; Creatify API credentials (a Pro plan or higher is required), API credentials can be obtained from the Creatify account settings. Installation: Installing from npm is recommended. Configuration: Set the Creatify API credentials as environment variables, or create a .env file. In Claude Desktop: Add the server to the Claude Desktop configuration (on macOS the path is ~/Library/Application Support/Claude/claude_desktop_config.json). It can also be used with a custom MCP client or run as a standalone server. After configuration is complete, natural language prompts can be used, for example: create an avatar video, convert a website into a promotional video, generate text-to-speech, list available avatars, check remaining credits, create a two-avatar conversation video, etc.
IV. Applicable Scenarios Suitable for AI assistant and automation tool scenarios that require automated video generation, including product demo videos, short social media videos (TikTok/Instagram/YouTube), educational tutorial videos, marketing promotional videos, video performance analysis, etc.
V. Limitations and Notes Creatify API credentials are required and the account must be on a Pro plan or higher; video creation usually takes 2-5 minutes; API response time is usually less than 2 seconds. Security recommendations: do not commit API keys to version control, use environment variables to store sensitive data, rotate API keys regularly, monitor API usage, and use HTTPS for all Webhook URLs. Performance optimization: avatar/voice lists can be cached for 1 hour, video status should be polled every 5-10 seconds for active tasks, and templates can be cached for 24 hours. The project uses the MIT license.
