MCP 服务器概览
what is open.video MCP?
open.video MCP is an AI-powered, MCP-compatible server that gives LLM agents full control of a video platform: upload and link videos, manage channels and playlists, generate embed codes, and fetch analytics — all via natural language through any MCP-capable AI client (Claude, Cursor, Windsurf, etc.). It is a remote-first, cloud-hosted server that mediates access to your publishing account.
how to use open.video MCP?
- Sign up for an open.video account and create a channel (no credit card required).
- Add the open.video MCP server to your client’s MCP config (example mcpServers JSON snippet).
- Optionally add the Upload Helper for local file uploads (npx open-video-upload-helper).
- Authenticate via OAuth in-browser (scoped tokens) and verify by calling list_channels.
- Use any MCP-compatible AI client or upload the provided LLM Skill to teach an LLM how to operate the platform.
key features
- OAuth browser-based authentication — no API keys required; tokens scoped per upload
- Remote-first cloud-hosted MCP server (no local install required)
- 16 MCP tools that cover Videos, Channels, Analytics, and Playlists (upload_video, link_video, get_embed_code, update_video, delete_video, list_videos, list_channels, get_channel, update_channel, upload_channel_image, get_analytics, get_video_status, create_playlist, update_playlist, add_to_playlist, list_playlists)
- Batch operations and instant commands via natural language (no scripts)
- Works with any MCP-compatible client and includes an LLM skill for deep automation
use cases
- Let an LLM agent upload, tag, and publish batches of videos and thumbnails
- Automate channel management and branding (update channel info, upload images)
- Generate embed codes and metadata for integration with websites or apps
- Track performance and engagement metrics via get_analytics and get_video_status
- Build, edit, and reorder playlists programmatically using natural language
- Integrate video operations into AI workflows for content teams, educators, or creators
FAQ
-
Q: How do I authenticate? A: OAuth browser sign-in with scoped tokens; no API keys are required.
-
Q: Which AI clients are supported? A: Any MCP-compatible client (examples: Claude, Cursor, Windsurf, Claude Desktop connectors).
-
Q: How long does setup take? A: The docs claim you can connect in under 2 minutes by adding the MCP server to your config and completing OAuth.
-
Q: Can I upload local files? A: Yes — use the Upload Helper (npx open-video-upload-helper) as a secondary MCP server for local file uploads.
-
Q: Is there an LLM integration or skill? A: Yes — a downloadable Open.Video LLM skill file is provided to teach LLMs how to use the platform.
-
Q: Do I need to self-host? A: No — open.video MCP is remote-first and cloud-hosted, so nothing needs to be installed locally unless you use the upload helper.
-
Q: What about pricing? A: Sign-up and channel creation note “no credit card required”; specific pricing plans are not detailed in the MCP overview — check the docs/website for current pricing.