MCP server · source category: Multimedia Process

haljishi/vidwords-mcp

Let an agent read a YouTube video and cite the exact second. Search one video's transcript for a topic and get back only…

Let an agent read a YouTube video and cite the exact second. Search one video's transcript for a topic and get back only the matching moments with clickable ?t= deep links, or analyze the frames — slides, charts, demos and on-screen text that is never spoken aloud. Every citation is checked against a recorded frame or a real transcript segment and dropped if it fails. Hosted Streamable HTTP at https://vidwords.com/mcp, token header or OAuth, free tier at vidwords.com.

Open documentation ↗

Record

Type
MCP server
Category
Text, Language & Speech
Authentication
API key
Pricing
Free

Fields are reproduced from the source registry as-is. “Unknown” means the source did not describe the field — it is not an inference. See the methodology for how records are collected and normalised.

Browse all Text, Language & Speech MCP servers.