# skribr MCP — setup & usage

skribr is a PAID MCP server (part of skribr Pro). It gives your AI agent context from any
video — YouTube, TikTok, Instagram, X, Facebook — and any web page or whole site, as
distilled, ready context (not raw transcripts).

## Requirements (paid)
- skribr Pro: $19/mo, includes 2,000 credits/month. 1 video or web page = 1 credit; site
  crawl and AI-extraction use more.
- An API key SKRIBR_API_KEY (format sk_skribr_…). You get it with Pro at https://skribr.net/mcp
  — without a valid key the server returns an upgrade link instead of results.

## Install — pick your client
Package: skribr-mcp  (runs via: npx -y skribr-mcp). Set env SKRIBR_API_KEY to your key.

Claude Code (one line):

    claude mcp add skribr --env SKRIBR_API_KEY=YOUR_KEY -- npx -y skribr-mcp

Codex CLI (one line):

    codex mcp add skribr --env SKRIBR_API_KEY=YOUR_KEY -- npx -y skribr-mcp

  or add to ~/.codex/config.toml (TOML, not JSON):

    [mcp_servers.skribr]
    command = "npx"
    args = ["-y", "skribr-mcp"]
    [mcp_servers.skribr.env]
    SKRIBR_API_KEY = "YOUR_KEY"

VS Code (GitHub Copilot) — one line (note: .vscode/mcp.json uses the "servers" key):

    code --add-mcp '{"name":"skribr","command":"npx","args":["-y","skribr-mcp"],"env":{"SKRIBR_API_KEY":"YOUR_KEY"}}'

Cursor, Windsurf, Gemini CLI, Cline, and most other MCP clients — same mcpServers block
(Cursor: ~/.cursor/mcp.json · Windsurf: ~/.codeium/windsurf/mcp_config.json ·
Gemini CLI: ~/.gemini/settings.json):

    {
      "mcpServers": {
        "skribr": {
          "command": "npx",
          "args": ["-y", "skribr-mcp"],
          "env": { "SKRIBR_API_KEY": "YOUR_KEY" }
        }
      }
    }

## Tools
- transcript — any video URL (YouTube/TikTok/Instagram/X/Facebook/file) -> transcript
- context — any video -> distilled agent context (agent_instruction / prd / study_guide / code)
- scrape — a web page -> clean markdown
- crawl — a whole site -> content
- map — a site -> all its URLs
- metadata — video / channel / playlist metadata
- extract — AI structured extraction from media
- enumerate — a channel / playlist -> list of videos

## Usage — just ask your agent
- Summarize this TikTok — tiktok.com/@user/video/…
- Crawl these docs and answer my question — example.com/docs
- Draft a PRD from this talk — youtu.be/…
- Pull the API endpoints from this page — example.com/api

Pro & full docs: https://skribr.net/mcp