MCP YouTube Transcript
An MCP server retrieving transcripts of YouTube videos
作者による説明(英語の原文)
Curated SISTRIX SEO tools: visibility, rankings, keywords, backlinks, AI visibility, Amazon data.
この3つを1つのスコアにまとめることはしません。意見、測定、人気はそれぞれ別のものだからです。
このサーバーのやさしい解説はまだ書いていません。以下はレジストリから読み取った事実です。
使っているAIアプリを選んで、設定をコピーしてください。
SISTRIX_API_KEY, SISTRIX_COUNTRY「秘密」と表示された値は、あなたのアカウントの鍵です。絶対に人に教えたり、チャットに貼り付けたりしないでください。
必要なもの
どのアプリを使っていますか?
上でアプリを選んでください。選んだアプリを覚えておくので、次からはボタンに「…にインストール」と表示されます。
Claude Desktop → Settings → Developer → Edit Config を開き、claude_desktop_config.json にこれを貼り付けます
{
"mcpServers": {
"sistrix": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"SISTRIX_API_KEY",
"-e",
"SISTRIX_COUNTRY",
"ghcr.io/liohtml/sistrix-mcp:v0.1.1"
],
"env": {
"SISTRIX_API_KEY": "your-value",
"SISTRIX_COUNTRY": "your-value"
}
}
}
}ターミナルでこれを実行します
claude mcp add sistrix --env SISTRIX_API_KEY=your-value --env SISTRIX_COUNTRY=your-value -- docker run -i --rm -e SISTRIX_API_KEY -e SISTRIX_COUNTRY ghcr.io/liohtml/sistrix-mcp:v0.1.1動かす前に、空欄の値を入力してください。
プロジェクトの .cursor/mcp.json、または ~/.cursor/mcp.json に貼り付けます
{
"mcpServers": {
"sistrix": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"SISTRIX_API_KEY",
"-e",
"SISTRIX_COUNTRY",
"ghcr.io/liohtml/sistrix-mcp:v0.1.1"
],
"env": {
"SISTRIX_API_KEY": "your-value",
"SISTRIX_COUNTRY": "your-value"
}
}
}
}ブラウザがアプリを開いてよいか確認し、そのあとアプリが設定を表示して確認を求めます。キーの値は空欄のままなので、自分で入力してください。
プロジェクトの .vscode/mcp.json に貼り付けます
{
"servers": {
"sistrix": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"SISTRIX_API_KEY",
"-e",
"SISTRIX_COUNTRY",
"ghcr.io/liohtml/sistrix-mcp:v0.1.1"
],
"env": {
"SISTRIX_API_KEY": "your-value",
"SISTRIX_COUNTRY": "your-value"
}
}
}
}ブラウザがアプリを開いてよいか確認し、そのあとアプリが設定を表示して確認を求めます。キーの値は空欄のままなので、自分で入力してください。
これはパソコンで動くサーバーですが、ChatGPTがつなげるのはインターネット上のサーバーだけです。ほかのアプリを選んでください。
選んだアプリは、この端末のCookieに保存されます。アカウントは不要です。
SISTRIX_COUNTRY envこの設定は、作者のレジストリ登録情報から作成しています。READMEと違う場合は、READMEに従ってください。
パソコンで動くサーバーです。インストールすると、ツール一覧がAIアプリに表示されます。
誰でも見られる情報をチェックしており、コードは読んでいません。当てはまらない項目は数に入れていません。 · 2026年11月6日にチェック
レジストリ名は、リポジトリを所有するGitHubアカウント Liohtml で確認済みです
このリポジトリはまだGitHubから読み取っていません
このリポジトリはまだGitHubから読み取っていません
これはパソコンで動くため、インターネット上のエンドポイントはありません
パソコンで動き、代わりにローカルの設定を使います
パソコンで動くサーバーです。ツール一覧を読み取れるのは、インターネット上のエンドポイントがある場合だけです
レジストリでこのカテゴリにある有効なすべてのサーバーと比較しています。スターは、公開者がリポジトリの所有者である場合だけ数えます。
当てはまる基本チェック1項目のうち1項目に合格しました。コードは読んでいません。キーやトークンを追加する前に、サーバーがどんなアクセスを求めているかを確認し、必要な権限だけを持つキーを使ってください。
入力が必要なもの:SISTRIX_API_KEY
あなたのパソコンで動きます。必要なときにAIアプリが起動し、データは作者のサーバーを通りません(サーバー自体がオンラインサービスを呼び出す場合を除きます)。
GitHubリポジトリの最終更新は–、最新バージョン(0.1.1)がレジストリに届いたのは4 か月前です。
Claude Desktop、Claude Code、Cursor、VS CodeなどのMCP対応アプリで使えます。パソコンにRust, Dockerも必要です。
同じカテゴリには:MCP YouTube Transcript, Google Analytics MCP, BibiGPT YouTube Video Summarizer
同じカテゴリで、ほかの作者が作ったサーバー。スターの多い順。
カードの見方:★ GitHubスターi3つの柱:レビュー · チェック · 利用状況i更新=最新バージョンをレジストリに公開した日i
An MCP server retrieving transcripts of YouTube videos
GA4 MCP server for AI agents: schema discovery, server-side aggregation, analysis-ready defaults.
AI video summarizer: summarize and transcribe YouTube, Bilibili, TikTok and podcast videos.
作者のリポジトリから書かれたとおりにコピーしたものです。PopNx MCPは内容を確認・推奨していません。HTML、画像、スクリプトは取り除いています。リンクはこのサイトの外へ移動します。
Talk to your SISTRIX SEO data. From Claude, Cursor, VS Code, or any MCP client.
*17 curated, credit-aware SEO tools + a full-API escape hatch. Single binary, instant startup, context-friendly.*
Quickstart · Clients · Tools · Credits · Configuration · FAQ
You ▸ How is example.com doing in Google — and who is beating us?
Claude ▸ example.com's visibility index is 4.82 (Germany), up 6% over the last 8 weeks. It ranks for 18,400 organic keywords, 2,100 in the top 10. Closest SEO competitors: competitor-a.de (74% overlap), competitor-b.com (61%). Biggest opportunity: 312 keywords sit on page 2 — want the top 20 by potential gain?
Every question the SISTRIX Toolbox can answer, your AI assistant can now answer too — including follow-ups, comparisons, and "why?".
| 🎯 Curated, not generated | 17 hand-crafted tools modeled on real SEO questions — not 90+ auto-generated API mirrors that flood the model's context and degrade tool selection. |
| 🧭 Guided workflows | Four MCP prompts (SEO health check, keyword research, competitor comparison, AI visibility report) run a complete credit-aware analysis with one click. |
| 💳 Credit-aware | SISTRIX bills most API calls per returned row from a weekly credit budget. Every tool ships tight row limits, cost notes in the descriptions, and a free sistrix_credits balance check — so one question never burns your week. |
| ⚡ Instant startup | No introspection round-trips. One static binary, no Node, no Python, no runtime. Starts in milliseconds. |
| 🔒 Safe by default | API key sent via POST body only (never in URLs/logs), redacted from every error message. |
| 🧠 Context-friendly | Row limits on every report and a hard response budget with actionable guidance — one tool call can never blow up the context window. |
| 🤖 AI visibility included | SISTRIX's newest data: which brands, entities, and sources ChatGPT, Perplexity, and Google AI Overviews actually cite (sistrix_ai_*). |
| 🛒 Amazon included | Marketplace data by ASIN: rankings, price history, reviews (sistrix_amazon). |
| 🧰 Never a cage | sistrix_api reaches any documented API method (ai.check, marketplace.keyword.*, project.onpage.*, …) when the curated tools don't cover it. |
| 🔁 Resilient | Automatic retries with backoff on 429/5xx/network hiccups. Every documented SISTRIX error code mapped to a hint the model can act on. |
Prebuilt binary (Linux, macOS, Windows) — grab it from Releases, or:
# Cargo
cargo install sistrix-mcp
cargo install --git https://github.com/Liohtml/sistrix-mcp
docker pull ghcr.io/liohtml/sistrix-mcp
app.sistrix.com/account/api → create a key. API access is included from the Plus package onwards and uses a weekly credit budget.
sistrix-mcp --api-key YOUR_KEY --check
✓ Connected — 10000 API credits available
Claude Code
claude mcp add sistrix \
--env SISTRIX_API_KEY=YOUR_KEY \ --env SISTRIX_COUNTRY=de \ -- sistrix-mcp
Claude Desktop
Add to claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):
{
"mcpServers": { "sistrix": { "command": "sistrix-mcp", "env": { "SISTRIX_API_KEY": "YOUR_KEY", "SISTRIX_COUNTRY": "de" } } } }
Cursor
.cursor/mcp.json (project) or ~/.cursor/mcp.json (global):
{
"mcpServers": { "sistrix": { "command": "sistrix-mcp", "env": { "SISTRIX_API_KEY": "YOUR_KEY", "SISTRIX_COUNTRY": "de" } } } }
VS Code (GitHub Copilot)
.vscode/mcp.json:
{
"servers": { "sistrix": { "type": "stdio", "command": "sistrix-mcp", "env": { "SISTRIX_API_KEY": "${input:sistrix-key}", "SISTRIX_COUNTRY": "de" } } }, "inputs": [ { "id": "sistrix-key", "type": "promptString", "description": "SISTRIX API key", "password": true } ] }
Windsurf / Zed / other MCP clients
Any client that speaks MCP over stdio works with the generic shape:
{
"command": "sistrix-mcp", "args": [], "env": { "SISTRIX_API_KEY": "YOUR_KEY", "SISTRIX_COUNTRY": "de" } }
Docker (any client)
{
"mcpServers": { "sistrix": { "command": "docker", "args": [ "run", "-i", "--rm", "-e", "SISTRIX_API_KEY", "-e", "SISTRIX_COUNTRY", "ghcr.io/liohtml/sistrix-mcp" ], "env": { "SISTRIX_API_KEY": "YOUR_KEY", "SISTRIX_COUNTRY": "de" } } } }
[!TIP]
Set
SISTRIX_COUNTRY(e.g.de) and every tool call automatically uses that
country index unless the model asks for another one.
| Tool | Answers questions like | Credits* |
|---|---|---|
sistrix_credits | *"How many API credits do we have left?"* | free |
sistrix_lists | *"Which country indices / SERP features / AI models exist?"* | free |
sistrix_domain_overview | *"How is example.com doing overall?"* | 5 flat |
sistrix_visibility | *"Visibility index trend over the last year? All-time high?"* | 1/row |
sistrix_keyword_counts | *"How many keywords do we rank for? Top-10 trend?"* | 1/row |
sistrix_competitors | *"Who are our closest SEO / Ads competitors?"* | 1/row |
sistrix_keyword_ideas | *"Which keywords are just off page 1 with the biggest potential?"* | 1/row |
sistrix_domain_structure | *"Which URLs/subdomains/directories drive our visibility?"* | 1/row |
sistrix_domain_rankings | *"All keywords we rank for on page 2, with traffic."* | 1/row |
sistrix_keyword | *"Volume, CPC, intent, SERP features for 'ergonomic chair'?"* | 1–5/row |
sistrix_links | *"How does our backlink profile look? Top link texts?"* | 1/row–25 flat |
sistrix_ai_top | *"Which brands does AI mention most in our market?"* | 1/row |
sistrix_ai_entity | *"How do ChatGPT & Co. see our brand? Which sources do they cite?"* | 1/row–10 flat |
sistrix_ai_tracker | *"How visible are we in our tracked AI prompts vs. competitors?"* | 1/row |
sistrix_project | *"Optimizer project rankings, onpage crawl results, tracked SERPs."* | varies |
sistrix_amazon | *"Amazon rankings, price and review history for this ASIN."* | 1/row |
sistrix_api | Everything else — ai.check.*, marketplace.keyword.*, project.onpage.*, any documented method | varies |
*Costs as documented by SISTRIX per method; see API credits.
Domain tools accept a target (the domain, host, path, or URL) plus a scope that tells SISTRIX how to interpret it, a country index, and a row limit.
Four ready-made analysis workflows ship as MCP prompts — clients like Claude surface them as one-click commands. Each runs a proven, credit-aware tool sequence and ends in a structured report:
| Prompt | Arguments | What you get |
|---|---|---|
seo_health_check | target, country? | Full status: key figures, trend, competitors, quick wins, next actions |
keyword_research | keyword, country? | Metrics, intent, SERP features, questions, who ranks — plus a verdict |
competitor_comparison | target, competitor, country? | Side-by-side comparison and 3 concrete moves to close the gap |
ai_visibility_report | brand, country? | How ChatGPT & Co. see the brand: competitors, prompts, cited sources |
SISTRIX API access starts with the Plus package and uses a weekly credit budget. Most methods cost 1 credit per returned row; a few are flat-rate (domain.overview = 5, domain.visibilityindex.overview = 10, links.overview = 25) and keyword.seo.metrics costs 5 per keyword.
sistrix-mcp defaults every list to 25 rows (not SISTRIX's 100–10,000) and tells the model what each call costs, so exploratory conversations stay cheap. Check the balance anytime with sistrix_credits — it's free, as are sistrix_lists and the project/tracker list reports.
The rate limit is 300 requests/minute; sistrix-mcp backs off and retries automatically when it's hit.
| Flag | Env | Default | Description |
|---|---|---|---|
--api-key | SISTRIX_API_KEY | — *(required)* | SISTRIX API key (create here) |
--country | SISTRIX_COUNTRY | account default | Default country index (e.g. de, at, us) for all country-aware tools |
--timeout-secs | SISTRIX_TIMEOUT_SECS | 30 | Per-request timeout |
--max-response-chars | SISTRIX_MAX_RESPONSE_CHARS | 50000 | Response budget before truncation |
--api-url | SISTRIX_API_URL | https://api.sistrix.com/ | API endpoint (testing only) |
--check | — | — | Verify the API key, print remaining credits, then exit |
SISTRIX documents its own MCP bridge for connecting chatbots to the API — and per their docs, requests through it currently don't count against the credit quota. If that covers your needs, use it!
sistrix-mcp is the open-source, self-hosted alternative: a local stdio binary with a curated, fully documented tool surface (17 tools with exact schemas, enums, and per-case validation), an escape hatch to every documented API method including the newest areas (ai.check, marketplace, Optimizer onpage), hard response budgets for the model's context, offline tests, and no dependency on a hosted bridge. MIT-licensed, so you can read exactly what the model can and cannot do with your API key.
"wrong api key" (code 100)
Run sistrix-mcp --api-key ... --check. If it fails, regenerate the key under app.sistrix.com/account/api.
"not enough credits" (code 200)
The weekly budget is exhausted. Credits refill weekly; check the balance with sistrix_credits. Keep limit small — most tools cost 1 credit per row.
</d
冒頭だけを表示しています。続きはリポジトリで読めます