vap-showcase
Media Execution Control Layer for AI Agents. Reserve-execute-burn/refund pattern. FFmpeg post-processing (format conversion, audio normalization) Supports Flux2 Pro, Veo 3.1, Suno V5.
claude mcp add --transport stdio vapagentmedia-vap-showcase python -m vap_showcase_server \ --env VAP_API_KEY="YOUR_API_KEY"
How to use
VAP ShowCase is an MCP server that exposes Flux-based image generation, Veo 3.1 video generation, and Suno V5 music generation through agent workflows. It provides cost-aware, deterministic media generation that can be driven from supported MCP clients such as Claude Desktop, Cursor, and VS Code adapters. Available tools include generate_image for text-to-image prompts, generate_video for cinematic video creation from prompts, and generate_music for original music generation. You can also perform post-processing and editing tasks like upscale, background_remove, inpaint, video_trim, and video_merge, as well as cost estimation and task status lookups (estimate_cost, get_task, list_tasks). The server enforces pre-commit pricing and a reserve-then-burn model so agents see predictable costs before execution and can manage budgets across tasks.
How to install
Prerequisites:
- Python 3.11+ installed on your system
- Access to an API key from vapagent (for authenticated operations)
- Internet access to install dependencies and reach the Vap API
Step-by-step installation:
-
Create a Python virtual environment (optional but recommended) python -m venv .venv source .venv/bin/activate # on macOS/Linux .venv\Scripts\activate # on Windows
-
Install required packages (adjust to your environment as needed) pip install -r requirements.txt # if a requirements file is provided by the project
or install directly if dependencies are listed in setup.py / pyproject.toml
-
Configure environment variables (example)
- VAP_API_KEY=your_api_key
- VAP_API_URL=https://api.vapagent.com
-
Run the MCP server python -m vap_showcase_server
-
Verify server is running
- Check the console for startup messages indicating the MCP endpoint (e.g., https://<host>/mcp)
- Use an MCP client (Claude Desktop, Cursor, or Cline) to connect with the provided API key
Optional: If you prefer running via a proxy or container, adapt the command to your environment but ensure VAP_API_KEY is passed as an env var.
Additional notes
Tips and common notes:
- Keep your VAP_API_KEY secure; rotate credentials as needed and avoid sharing plaintext keys in configs.
- The MCP server exposes a set of tools (generate_image, generate_video, generate_music, upscale, background_remove, inpaint, video_trim, video_merge, estimate_cost, check_balance, get_task, list_tasks). Use the SDK or MCP clients to invoke these tools from your agent workflow.
- For local proxies or environments that don’t support headers, you can use the local proxy approach described in the integration docs, pointing the proxy to https://api.vapagent.com/mcp and supplying the API key via environment variables.
- If you run into rate limits or cost estimation discrepancies, verify that pre-commit pricing and reserve-burn behavior are enabled and that the correct currency/units are configured in your account.
- The registry name for this server is vap-e (as referenced in the MCP Registry).
Related MCP Servers
openapi
OpenAPI definitions, converters and LLM function calling schema composer.
comfy-pilot
MCP server + embedded terminal that lets Claude Code see and edit your ComfyUI workflows
seedream-image
🚀 PixelMCP | 为你的 Cursor、Claude Code 等集成AI绘画能力,让AI生成的页面不再单调!
skill-to
Convert AI Skills (Claude Skills format) to MCP server resources - Part of BioContextAI
obsidian
MCP server for Obsidian vault management - enables Claude and other AI assistants to read, write, search, and organize your notes
nanobanana
Gemini Vision & Image Generation MCP for Claude Desktop and Claude Code