mcp-youtube-transcript
MCP YouTubeトランスクリプトサーバー
YouTube動画からトランスクリプトを取得できるモデルコンテキストプロトコルサーバーです。シンプルなインターフェースから動画のトランスクリプトに直接アクセスできるため、コンテンツの分析や処理に最適です。
目次
Related MCP server: YouTube Transcript Server
特徴
✨ 主な機能:
YouTube動画からトランスクリプトを抽出する
複数の言語をサポート
連続モードまたは段落モードでテキストをフォーマットする
ビデオのタイトルとメタデータを取得する
自動段落分割
テキストの正規化とHTMLエンティティのデコード
堅牢なエラー処理
タイムスタンプと重複検出
はじめる
前提条件
Node.js 18以上
インストール
インストール方法は 2 つあります。
オプション 1: 手動構成 (本番環境に推奨)
Claude Desktop 構成ファイルを作成または編集します。
macOS:
~/Library/Application Support/Claude/claude_desktop_config.jsonWindows:
%APPDATA%\Claude\claude_desktop_config.json
次の構成を追加します。
{
"mcpServers": {
"youtube-transcript": {
"command": "npx",
"args": [
"-y",
"@sinco-lab/mcp-youtube-transcript"
]
}
}
}macOS のクイックセットアップ スクリプト:
# Create directory if it doesn't exist
mkdir -p ~/Library/Application\ Support/Claude
# Create or update config file
cat > ~/Library/Application\ Support/Claude/claude_desktop_config.json << 'EOL'
{
"mcpServers": {
"youtube-transcript": {
"command": "npx",
"args": [
"-y",
"@sinco-lab/mcp-youtube-transcript"
]
}
}
}
EOLオプション 2: Smithery 経由 (開発のみ)
npx -y @smithery/cli install @sinco-lab/mcp-youtube-transcript --client claude⚠️注意: この方法は Smithery のプロキシ サービスに依存しているため、本番環境での使用には推奨されません。
使用法
基本構成
Claude Desktop / Cursor / cline で使用するには、構成が一致していることを確認してください。
{
"mcpServers": {
"youtube-transcript": {
"command": "npx",
"args": ["-y", "@sinco-lab/mcp-youtube-transcript"]
}
}
}テスト
クロード・アプ氏と
インストール後にClaudeアプリを再起動します
簡単なコマンドでテストします。
https://www.youtube.com/watch?v=AJpK3YTTKZ4 Summarize this video
出力例:
MCPインスペクターを使用
# Clone and setup
git clone https://github.com/sinco-lab/mcp-youtube-transcript.git
cd mcp-youtube-transcript
npm install
npm run build
# Launch inspector
npx @modelcontextprotocol/inspector node "dist/index.js"
# Access http://localhost:6274 and try these commands:
# 1. List Tools: clink `List Tools`
# 2. Test get_transcripts with:
# url: "https://www.youtube.com/watch?v=AJpK3YTTKZ4"
# lang: "en" (optional)
# enableParagraphs: false (optional)トラブルシューティングとメンテナンス
クロードログの確認
Claude のログを監視するには、次のコマンドを使用できます。
tail -n 20 -f ~/Library/Logs/Claude/mcp*.logこれにより、ログ ファイルの最後の 20 行が表示され、新しいエントリが追加されるたびに表示され続けます。
注:ClaudeアプリはMCPサーバーのログファイルに自動的に
mcp-server-プレフィックスを付けます。例えば、サーバーのログはmcp-server-youtube-transcript.logに書き込まれます。
npxキャッシュのクリーニング
npxキャッシュに関連する問題が発生した場合は、次のコマンドを使用して手動でクリーンアップできます。
rm -rf ~/.npm/_npxこれにより、キャッシュされたパッケージが削除され、新しく開始できるようになります。
APIリファレンス
トランスクリプトを取得する
YouTube 動画からトランスクリプトを取得します。
パラメータ:
url(文字列、必須): YouTube 動画の URL または IDlang(文字列、オプション): 言語コード (デフォルト: "en")enableParagraphs(ブール値、オプション):段落モードを有効にする(デフォルト:false)
応答形式:
{
"content": [{
"type": "text",
"text": "Video title and transcript content",
"metadata": {
"videoId": "video_id",
"title": "video_title",
"language": "transcript_language",
"timestamp": "processing_time",
"charCount": "character_count",
"transcriptCount": "number_of_transcripts",
"totalDuration": "total_duration",
"paragraphsEnabled": "paragraph_mode_status"
}
}]
}発達
プロジェクト構造
├── src/
│ ├── index.ts # Server entry point
│ ├── youtube.ts # YouTube transcript fetching logic
├── dist/ # Compiled output
└── package.json主要コンポーネント
YouTubeTranscriptFetcher: コアトランスクリプト取得機能YouTubeUtils: テキスト処理とユーティリティ
機能と性能
エラー処理:
無効なURL/ID
利用できないトランスクリプト
言語の可用性
ネットワークエラー
レート制限
テキスト処理:
HTMLエンティティのデコード
句読点の正規化
空間正規化
スマートな段落検出
貢献
貢献を歓迎します!問題やプルリクエストをお気軽にご提出ください。
ライセンス
このプロジェクトは MIT ライセンスに基づいてライセンスされています - 詳細についてはLICENSEファイルを参照してください。
関連プロジェクト
Available Tools
1 toolget_transcriptsA
Extract and process transcripts from a YouTube video.
Parameters:
url(string, required): YouTube video URL or ID.lang(string, optional, default 'en'): Language code for transcripts (e.g. 'en', 'uk', 'ja', 'ru', 'zh').enableParagraphs(boolean, optional, default false): Enable automatic paragraph breaks.
IMPORTANT: If the user does not specify a language code, DO NOT include the lang parameter in the tool call. Do not guess the language or use parts of the user query as the language code.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | YouTube video URL or ID | |
| lang | No | Language code for transcripts, default 'en' (e.g. 'en', 'uk', 'ja', 'ru', 'zh') | en |
| enableParagraphs | No | Enable automatic paragraph breaks, default `false` |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It effectively explains the tool's core function and includes important behavioral guidance about parameter handling (the IMPORTANT note about not guessing language). However, it doesn't mention potential limitations like video availability, transcript existence, rate limits, or error conditions.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured with a clear purpose statement followed by parameter documentation and important usage notes. Every sentence serves a purpose, though the parameter list slightly duplicates schema information. The IMPORTANT section is appropriately emphasized for critical behavioral guidance.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no annotations and no output schema, the description provides adequate coverage for the tool's basic function and parameters. However, it lacks information about return values, error handling, and operational constraints that would be helpful for an agent. The IMPORTANT note adds valuable context, but more behavioral transparency would improve completeness.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the schema already documents all parameters thoroughly. The description repeats this information in a bulleted list without adding significant semantic context beyond what's in the schema. The IMPORTANT note about language parameter handling adds some value, but overall the description doesn't enhance parameter understanding beyond the structured schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose with specific verbs ('extract and process') and resource ('transcripts from a YouTube video'). It distinguishes itself from potential alternatives by focusing on transcript extraction rather than other video-related operations, though no sibling tools exist for direct comparison.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides some usage guidance through the IMPORTANT note about language parameter handling, but it doesn't explicitly state when to use this tool versus alternatives (e.g., when transcripts are needed vs. other video metadata). Since no sibling tools exist, this is less critical, but general context about appropriate use cases is missing.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
1 tool update
- First observed
get_transcripts
TDQS
Scored across 1 tool
With only one tool, there is no possibility of ambiguity or overlap between tools. The tool's purpose is clearly defined and distinct by default.
Since there is only one tool, naming consistency is inherently perfect. The tool name 'get_transcripts' follows a clear verb_noun pattern.
A single tool is too few for a server focused on YouTube transcripts, as it lacks operations like searching transcripts, managing multiple videos, or handling errors. This minimal scope limits functionality and agent workflows.
The tool set is severely incomplete for the domain of YouTube transcript processing. It only provides extraction (get_transcripts), missing essential operations such as searching within transcripts, listing available languages, or handling video metadata, which are common needs in this context.
Maintenance
Related MCP Connectors
An MCP server that gives any LLM or agent clean YouTube transcripts on demand: a single video, a whole channel, or a playlist, plus AI cleanup of auto-generated captions. API-key auth, credit-based, same backend as the public v1 API. Get a free API key with 25 free credits at youtubetranscriptdownload.com/account.
YouTube transcripts, search, channel browsing, and playlists for AI agents via MCP.
MCP server for RiverScript, an AI transcription platform - fetches transcripts shared via a link.
Transcripts of YouTube videos, playlists and channels in any language: text, SRT, VTT or JSON.
Related MCP Servers
- AlicenseAqualityFmaintenanceA Model Context Protocol server that enables retrieval of transcripts from YouTube videos. This server provides direct access to video captions and subtitles through a simple interface.1656 npm595MIT
- AlicenseBqualityDmaintenanceA Model Context Protocol server that enables retrieval of transcripts from YouTube videos with language-specific support.1656 npm1MIT
- AlicenseBqualityDmaintenanceA Model Context Protocol server that enables AI assistants to extract transcripts from YouTube videos, allowing AI to analyze and work with video content directly.111 npm4MIT
- AlicenseAqualityDmaintenanceA Model Context Protocol server that enables access to YouTube video content through transcripts, translations, summaries, and subtitle generation in various languages.55MIT