vidcap

The VidCap MCP Server acts as a proxy layer to the VidCap YouTube API, exposing a set of MCP tools that let AI assistants query YouTube data and perform common tasks such as retrieving video info, discovering available media formats, fetching captions, generating AI-assisted summaries, capturing screenshots, and querying comments and search results. The server is designed to be consumed by any MCP-compatible assistant, with a focus on ease-of-setup via npx. To use the server, set your VidCap API key in the environment (VIDCAP_API_KEY) and start the server through your MCP client or by running the provided npx command. The server exposes endpoints under the VidCap YouTube API path and also provides eight MCP tools that can be invoked by your assistant for end-to-end video data access and analysis.

Available MCP tools include: get video info (getInfo), get media formats (getMedia), get captions (getCaption), get AI-generated summary (getSummary), get a single screenshot (getScreenshot), get multiple screenshots (getScreenshotMultiple), get comments (getComments), and search (youtube_search) for YouTube videos with advanced filtering. You can drive these tools via CLI examples described in the README or integrate them into your MCP client configuration for seamless prompts-to-queries execution.

Prerequisites:

• Node.js (v14+ recommended) and npm/yarn installed
• Access to the internet to fetch the MCP server package
• VidCap API key to be set in your environment (VIDCAP_API_KEY)

Installation steps:

1. Ensure Node.js is installed. Verify: node -v npm -v

2. Use npx to run the VidCap MCP server directly (no global install required): npx -y vidcap-mcp-server

   This will start the MCP server in your current environment. If you prefer a persistent startup, you can also preload it via a local or container-based workflow.

3. Set up environment variables (example for POSIX shells): export VIDCAP_API_KEY="your-vidcap-api-key"

4. Alternatively, if you manage via a process manager or container, configure the same environment variable inside that context and start the server through your usual startup script.

5. Verify the server is reachable via its MCP endpoints as described in the documentation (CLI examples or MCP tool invocations).

Note: The README references running via npm scripts (e.g., npm run dev:server:http) in a project context. If you clone the repository, you may install dependencies and use the provided npm scripts as documented there. The npx approach is package-based and requires no global installation.

Environment variables and configuration:

• VIDCAP_API_KEY: Required to access VidCap YouTube data. Ensure this is kept secret and loaded into environments where the MCP server runs.
• If you encounter rate limits or authentication errors, verify your API key and consider caching strategies described in the endpoints (e.g., the info and summary endpoints support optional caching).

Transport and usage tips:

• The server supports stdio (CLI) and Streamable HTTP transports for web clients. If you plan to use web-based MCP clients, ensure your host can reach the server via the HTTP transport endpoints.
• The MCP tools are designed to be invoked by an MCP-compatible assistant. For example, you can issue tool calls like youtube.getInfo or youtube.getCaption through your agent, providing required parameters (url, etc.).

Common issues:

• Missing VIDCAP_API_KEY: The server will not be able to fetch data from VidCap; set the environment variable.
• Network or DNS problems: Ensure the host running the MCP server can reach VidCap endpoints.
• Deprecated transport notes: The sse transport is deprecated; prefer stdio or streamable-http for compatibility.

Configuration options:

• You can adjust tool parameters (locale, model, cache, etc.) per endpoint call to tailor responses to your use case.