Provides tools for interacting with YouTube's Data API, enabling video and channel research, transcript extraction, audience sentiment analysis, viral outlier detection, SEO metadata generation, and creator monetization strategy.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@YouTube MCP Server Analyze the last 5 videos from MrBeast for SEO tags"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
33 Tools · 11 Autonomous Agent Skills · Zero-Quota RSS Fallback · Multi-Key Pool · 101 Tests
YouTube MCP Server connects your AI assistants — Claude Desktop, Cursor, Windsurf, Antigravity, and any MCP-compatible LLM agent — directly to live YouTube data.
Unlike basic toy wrappers that only search videos, this is an autonomous YouTube growth agency in a box: reverse-engineer competitors, detect 3x+ viral outliers, extract transcripts without an API key, generate high-CTR thumbnail prompts, hunt paying brand sponsors, and execute complete 7-phase channel strategies.
npx -y @smithery/cli install youtube-mcp --client claudeOr manually add to your claude_desktop_config.json:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"youtube": {
"command": "uvx",
"args": ["youtube-mcp"],
"env": {
"YOUTUBE_API_KEY": "YOUR_GOOGLE_CLOUD_API_KEY"
}
}
}
}Add to your workspace .cursor/mcp.json or global settings:
{
"mcpServers": {
"youtube": {
"command": "uvx",
"args": ["youtube-mcp"],
"env": {
"YOUTUBE_API_KEY": "YOUR_GOOGLE_CLOUD_API_KEY"
}
}
}
}# Instant run with uvx (no installation needed)
export YOUTUBE_API_KEY="your_key_here"
uvx youtube-mcp
# Or install via pip
pip install youtube-mcp
youtube-mcpRelated MCP server: YouTube MCP Server
Feature | YouTube MCP Server | Standard YouTube Wrappers |
33 Specialized Creator Tools | ✅ Full Suite | ❌ 2–3 basic endpoints |
Multi-Key Quota Rotation Pool ( | ✅ Zero 403 Crashes | ❌ Hard quota failures |
Zero-Quota RSS Feed ( | ✅ Free public Atom feed | ❌ Burns 100 quota units |
Transcript Extraction Without API Key | ✅ Instant subtitles | ❌ Requires OAuth |
Viral Outlier Detection (2.5x–10x spikes) | ✅ Built-in | ❌ None |
Sponsor Discovery & Rate Card Pricing | ✅ Automated CPM cards | ❌ None |
11 Turnkey AI Agent Skills | ✅ 7-phase autonomous pipeline | ❌ None |
Token-Optimized Markdown Payloads | ✅ Designed for LLM context | ❌ Raw bloated JSON dumps |
Strict 100% Live Data Guarantee | ✅ Authentic YouTube data only | ❌ Static/fake mocks |
Automated Test Coverage | ✅ 101 Unit & Integration Tests | ❌ Untested |
Every tool strips away raw JSON noise and returns structured, high-signal Markdown payloads optimized for LLM reasoning and minimal context usage.
Tool | Quota Cost | Description |
| 100 | Search YouTube with date filters, relevance, region codes, and duration. |
| 100 | Discover niche channels enriched with subscriber counts and total views. |
| Multi | Multi-niche creator scouting with subscriber range and activity filters. |
| Multi | Uncover emerging creators (1k–50k subs) with 3x+ engagement velocity. |
| 1 | Fetch real-time trending videos by country and video category. |
| 0 (Free) | Zero-Quota: Pull recent channel uploads via public RSS with no API key. |
Tool | Quota Cost | Description |
| 0 (Free) | Extract video captions and transcripts. Does not require an API key. |
| 1 | Retrieve structured video stats, tags, descriptions, and duration. |
| 1 | Pull top-level comments and discussions for viewer sentiment analysis. |
| 1 | Classify viewer comments into positive/negative sentiment, FAQs, and requests. |
| 0 (Free) | Identify punchy 30–60 second viral clip segments from transcripts. |
| 0 (Free) | Automatically generate YouTube description timestamps and chapter titles. |
Tool | Quota Cost | Description |
| 1 | Channel stats, subscriber count, branding, and uploads playlist ID. |
| 1 | List videos within any playlist or channel upload stream. |
| Multi | Full channel diagnosis: upload cadence, view consistency, and health score. |
| Multi | Identify videos generating 2.5x–10x the channel's baseline view count. |
| Multi | Head-to-head benchmarking of 2–5 competing channels across all metrics. |
| Multi | Measure how Shorts production impacts long-form subscriber conversion. |
| Multi | Generate upload frequency heatmaps and identify competitor gaps. |
Tool | Quota Cost | Description |
| Multi | Find high-demand search topics where existing videos have low view counts. |
| Multi | Predict whether a topic will succeed primarily via Search or Browse/Suggested. |
| Multi | Craft 10 psychological title formulas (curiosity, fear of missing out, extremes). |
| Multi | 3 counter-positioned thumbnail visual concepts with Midjourney/DALL-E prompts. |
| Multi | 7-beat retention script structure (0–30s hook, value bridge, payoff loop). |
| Multi | Complete upload kit: CTR title, description, tags, and chapter markers. |
| Multi | Structure 5-video series playlists to maximize Session Watch Time. |
| 0 (Free) | Calculate Words-Per-Minute (WPM) pacing to spot retention dropoff zones. |
Tool | Quota Cost | Description |
| Multi | Detect active brand sponsors in video descriptions across any niche. |
| Multi | Compute data-backed CPM and flat-rate sponsorship pricing. |
| Multi | Generate personalized brand outreach emails highlighting engagement metrics. |
| Multi | Build high-margin revenue funnels for channels under 1,000 subscribers. |
| Multi | Find high-view English videos with zero competition in Spanish/Portuguese/German. |
| Multi | Create interactive Community Tab polls that drive algorithmic re-engagement. |
The repository includes a ready-to-use Antigravity Plugin (.agents/plugins/youtube-creator-suite/) and 11 self-contained skills in .agents/skills/.
An AI agent can chain all tools into an end-to-end strategy automatically:
graph LR
A["Reconnaissance"] --> B["Strategy & Gaps"]
B --> C["Script & SEO"]
C --> D["Monetization"]
D --> E["Executive Dossier"]Skill | Slash Command | What It Delivers |
Full Auto Pipeline |
| 🔄 End-to-End Orchestrator: Chains all 10 skills in 7 automated phases from a single niche keyword. |
Channel Growth Audit |
| Diagnoses channel health, retention, and viewer sentiment with a 90-day turnaround plan. |
Viral Video Ideation |
| Content gap mining, title CTR psychological scoring, and Midjourney thumbnail prompts. |
Competitor Intelligence |
| Benchmarks rival channels head-to-head and extracts active paying sponsors. |
Channel Launch Architect |
| 0-to-1 launch blueprint: 5-video binge loops, 7-beat retention scripts, and upload packs. |
Content Repurposing |
| Transcribes longform videos into timestamped chapters and viral Shorts clips. |
Sponsor Pitch Architect |
| Discovers niche sponsors, calculates CPM rate cards, and drafts cold pitch emails. |
Publishing Scheduler |
| Analyzes competitor upload times to find low-competition Sweet Spot windows. |
International Arbitrage |
| Uncovers zero-competition opportunities in non-English markets with localized tags. |
Breakout Creator Scout |
| Scouts micro-influencers (1k–50k subs) with 3x+ velocity for partnership rosters. |
Executive Dossier |
| Compiles institutional-grade research dossiers formatted in Markdown or Notion. |
Once connected, simply ask your assistant in natural language:
Channel Turnaround:
"Audit @veritasium's channel. Identify their top viral outliers, analyze their upload cadence, and give me 3 video concepts to out-compete them."
Viral Packaging:
"Find content gaps in the 'AI automation' niche. Generate 5 high-CTR title formulas and 3 counter-positioned thumbnail visual concepts with Midjourney prompts."
Transcript Repurposing:
"Extract the transcript from this video URL https://youtu.be/... and write 3 high-hook YouTube Shorts scripts with visual cues."
Sponsor Outreach:
"Find which brand sponsors are actively spending money in the 'productivity' niche, compute a rate card for a 50k subscriber channel, and write a pitch email."
Complete Strategy:
"/youtube-full-pipeline niche='b2b saas growth' competitor_handles=['@saasstr', '@microconf']"
Create a .env file in your working directory (or set environment variables in your MCP client config):
# Required for YouTube Data API v3 queries
YOUTUBE_API_KEY="AIzaSy..."
# Optional: Multi-key quota failover pool (comma-separated)
YOUTUBE_API_KEYS="key_1,key_2,key_3"
# Optional: Server transport settings (stdio, sse, streamable-http)
MCP_TRANSPORT="stdio"
MCP_PORT=8000
MCP_HOST="127.0.0.1"
# Optional: Disk cache TTL in seconds (default: 3600)
MCP_CACHE_TTL=3600No API Key? Tools like get_channel_rss_videos, get_video_transcript, extract_shorts_clips, and analyze_transcript_pacing operate via public feeds and transcript streams with zero API quota requirements.
Run as an isolated, containerized remote MCP server:
# Build and run container
docker compose up -d
# Check health endpoint
curl http://localhost:8000/healthContributions are welcome!
📖 Review CONTRIBUTING.md to set up your local development environment.
🧪 Run tests with pytest (101 automated tests ensuring zero regression).
🔒 See SECURITY.md for vulnerability reporting.
💡 Open an Issue or join discussions to request new creator tools.
MIT License — free for personal, commercial, and agency use.
Mine video comments to discover audience pain points, questions, and content requests.
Identifies what viewers struggle with, questions the creator missed, topics requested for future videos, and the highest-upvoted comments.
Args: video_id_or_url: YouTube Video ID or full URL. max_comments: Number of top comments to analyze (up to 100, default 100).
Inputs: video_id_or_url (required), max_comments. Returns: result
Generate high-engagement Community Tab polls, quizzes, and discussion posts.
Leverages YouTube's Community Tab algorithm, which distributes polls into the home feeds of non-subscribers, creating viral discovery for new channels between video releases.
Args: niche_or_channel: Topic, niche, or creator handle (e.g. 'python programming', 'personal finance'). target_goal: Goal for the community strategy ('growth', 'video_validation', 'audience_loyalty').
Inputs: niche_or_channel (required), target_goal. Returns: result
Analyze competitor publishing schedules to find the optimal day and hour to upload.
Inspects publishing timestamps of top videos in a niche or channel, builds a day-of-week and hour-of-day distribution, and identifies low-competition "Sweet Spot" windows.
Args: niche_or_channel: Niche topic keyword (e.g. 'coding tutorials', 'finance') or creator handle. sample_size: Number of recent competitor uploads to sample (10 to 50, default 25). timezone_offset_hours: Timezone offset from UTC in hours (e.g. -5 for EST, +1 for CET, default 0).
Inputs: niche_or_channel (required), sample_size, timezone_offset_hours. Returns: result
Analyze a channel's balance between YouTube Shorts and Long-Form videos.
Calculates publishing ratio, view disparities, conversion efficiency, and provides a customized publishing mix recommendation to avoid Shorts cannibalizing long-form watch time.
Args: channel_id_or_handle: Channel handle (e.g. '@aliabdaal', '@mkbhd') or Channel ID. sample_videos: Number of recent uploads to evaluate (10 to 50, default 20).
Inputs: channel_id_or_handle (required), sample_videos. Returns: result
Reverse-engineer any YouTube channel's content strategy and monetization model.
Analyzes upload cadence, view performance distribution, title formulas, monetization links (sponsors, newsletters, affiliate links), and extracts the opening script hook from their top-performing video.
Args: channel_id_or_handle: Channel handle (e.g. '@mkbhd', '@aliabdaal'), channel ID ('UC...'), or username. sample_videos: Number of recent uploads to analyze (3 to 15, default 5).
Inputs: channel_id_or_handle (required), sample_videos. Returns: result
Conduct a full market research study and launch blueprint for starting a new YouTube channel.
Specifically engineered for beginners and new creators. Validates topic demand, identifies 3-5 realistic mid-sized channels to model (1k-300k subs), discovers proven outlier video topics where smaller channels blew up, mines audience pain points from comments, and provides a ready-to-record 5-video launch roadmap with recommended upload cadence.
Args: niche: The niche/topic you want to launch a channel in (e.g. 'ai automation', 'budget travel', 'personal finance for teens'). target_audience: Optional description of who your videos are for (e.g. 'complete beginners', 'busy moms'). region_code: Geographic market code (default 'US', e.g. 'GB', 'CA', 'DE').
Inputs: niche (required), region_code, target_audience. Returns: result
Classify whether a video topic will succeed via Evergreen Search or Viral Browse Feeds.
Provides algorithmic traffic predictions, expected RPM / AdSense monetization multipliers, longevity expectations (3+ years vs 14 days), and optimized title variants for both traffic channels.
Args: topic_or_title: Candidate video title, topic, or draft concept. target_niche: Optional niche context (e.g. 'coding', 'personal finance', 'fitness').
Inputs: topic_or_title (required), target_niche. Returns: result
Perform a head-to-head benchmarking comparison between 2 to 10 competing YouTube channels.
Compares subscriber growth velocity, recent average views, upload cadence, views-to-subscriber efficiency, and highlights category winners.
Args: channel_handles: List of 2 to 10 channel handles (e.g. ['@channel1', '@channel2']) or Channel IDs.
Inputs: channel_handles (required). Returns: result
Architect a 4-to-6 video binge-watching loop engineered to trigger YouTube's Session Watch Time multiplier.
Structures interconnected video concepts with seamless cliffhanger bridges and end-screen scripts so viewers watch multiple videos in sequence, signalling algorithmic promotion.
Args: core_topic: Overarching topic or learning journey (e.g. 'Build a SaaS in Python', 'Notion for Beginners'). video_count: Number of videos in the binge playlist series (3 to 6, default 5). target_audience: Optional target audience context.
Inputs: core_topic (required), video_count, target_audience. Returns: result
Discover brands actively sponsoring creators in a niche with discount codes and URLs.
Reveals companies with active influencer marketing budgets in your niche so you can pitch them as soon as you reach 1k-5k views per video.
Args: niche_or_query: Niche topic or keyword (e.g. 'productivity apps', 'coding', 'fitness'). sample_videos: Number of top videos to inspect (10 to 30, default 20). region_code: Country code (default 'US').
Inputs: niche_or_query (required), region_code, sample_videos. Returns: result
Generate a publication-ready Markdown research report for Notion or Obsidian.
Runs comprehensive market research and formats it into an executive-ready .md document complete with tables, competitor rankings, viral video links, comment insights, and a 5-video roadmap.
Args: niche: Topic or niche (e.g. 'ai automation', 'productivity systems', 'personal finance'). target_audience: Optional target audience description (e.g. 'beginners', 'freelancers'). output_file: Optional path where to save the markdown file (defaults to reports/_research_report.md). region_code: Country market code (default 'US').
Inputs: niche (required), output_file, region_code, target_audience. Returns: result
Generate YouTube chapters and extract the top 30-60s viral Shorts/Reels clips from transcripts.
Does NOT require a YouTube API key. Scans spoken transcripts for psychological hook triggers and formats timestamps ready to paste into video descriptions or feed into video editors.
Args: video_id_or_url: YouTube Video ID or full URL. min_duration_seconds: Minimum clip duration for Shorts/Reels (default 20). max_duration_seconds: Maximum clip duration for Shorts/Reels (default 60). max_clips: Maximum number of candidate clips to return (default 5).
Inputs: video_id_or_url (required), max_clips, max_duration_seconds, min_duration_seconds. Returns: result
Find modern breakout channels created recently that grew rapidly from scratch.
Discovers new channels that recently solved the YouTube algorithm rather than legacy giants.
Args: niche: Topic or niche keyword (e.g. 'ai automation', 'finance beginners'). max_channel_age_months: Maximum channel age in months (default 24). min_subscribers: Minimum subscribers (default 1000). max_subscribers: Maximum subscribers (default 300000). region_code: ISO country code (default 'US'). max_results: Max results to return (default 10).
Inputs: niche (required), max_results, region_code, max_subscribers, +2 more. Returns: result
Identify high-demand content gaps and low-competition keyword opportunities.
Finds topics where top search results are outdated (2+ years old), signaling easy ranking opportunities for a new channel to displace them with a fresh 2026 update.
Args: niche_or_topic: Topic, query, or question to analyze (e.g. 'how to learn sql for data analysis'). max_results: Number of search results to inspect (default 15). region_code: ISO country code (default 'US').
Inputs: niche_or_topic (required), max_results, region_code. Returns: result
Identify proven viral US/English video concepts with low competition in non-English markets.
Analyzes viral performance of English videos and checks competition levels in the target language / region (Spanish, French, German, Portuguese, Italian, Arabic, Japanese), providing translated title frameworks and market arbitrage scores.
Args: topic: Core topic in English (e.g. 'notion for students', 'ai automation', 'intermittent fasting'). target_language: Target language code ('es', 'fr', 'de', 'pt', 'it', 'ar', 'ja'). target_region: Target country code ('ES', 'MX', 'FR', 'DE', 'BR', 'IT', 'JP', 'SA'). max_results: Max opportunities to return (default 5).
Inputs: topic (required), max_results, target_region, target_language. Returns: result
Identify viral outlier videos that perform dramatically above a channel's normal average.
Discovers breakout video ideas from either a specific creator (e.g. '@creator') or across a broad niche topic (e.g. 'productivity tools') by flagging videos whose views surpass the channel's subscriber/view baseline by 2.5x, 5x, or 10x+.
Args: query: Niche topic keyword or creator handle ('@creator'). min_multiplier: Outlier threshold multiplier (default 2.5 means 2.5x the channel benchmark). published_after: RFC 3339 datetime to filter recent breakouts (e.g. '2024-01-01T00:00:00Z'). max_results: Number of candidates to evaluate (up to 50, default 25).
Inputs: query (required), max_results, min_multiplier, published_after. Returns: result
Architect 3 high-converting day-one monetization offers for channels with under 1,000 subscribers.
Enables creators to generate $500 - $3,000/mo from digital products, lead magnets, and consulting without waiting to reach the YouTube Partner Program AdSense threshold.
Args: niche: Topic or niche (e.g. 'notion productivity', 'python coding', 'budget travel'). target_audience: Target viewer demographic (e.g. 'freelancers', 'students', 'beginners'). main_skill_or_topic: Specific core skill being taught (optional).
Inputs: niche (required), target_audience, main_skill_or_topic. Returns: result
Generate a full 8-12 minute retention-engineered YouTube video script outline.
Reverse-engineers competitor transcripts for opening hooks and integrates real viewer pain points from comments to maximize watch time and viewer satisfaction.
Args: video_title_or_topic: The topic or title of the video to outline. competitor_video_id_or_url: Optional competitor video to model hook and structure from. target_audience: Ideal viewer demographic (default 'Beginners'). target_duration_minutes: Target video runtime in minutes (default 10).
Inputs: video_title_or_topic (required), target_audience, target_duration_minutes, competitor_video_id_or_url. Returns: result
Generate a complete YouTube Studio upload package: 3 mobile titles, chapters, description, tags, and pinned comment.
Args: topic: Primary topic or draft title of the video. key_takeaways: Optional list of main points covered in the video. channel_name: Optional creator channel name. affiliate_links: Optional list of affiliate/product URLs.
Inputs: topic (required), channel_name, key_takeaways, affiliate_links. Returns: result
Generate 3 distinct high-CTR thumbnail visual concepts with AI image prompts.
Deconstructs title-to-thumbnail contrast, visual color theory, facial expressions, and provides ready-to-use prompts for Midjourney, DALL-E, or Imagen.
Args: video_title: The title or topic of the video (e.g. 'How I Built a $10k/mo Micro-SaaS'). target_niche: Optional niche context (e.g. 'coding', 'business', 'productivity'). competitor_video_id_or_url: Optional competitor video to evaluate thumbnail benchmarks.
Inputs: video_title (required), target_niche, competitor_video_id_or_url. Returns: result
Retrieve channel information, subscriber counts, total views, and uploads playlist ID.
Provide at least one identifier: channel_id (e.g. 'UC...'), for_handle (e.g. '@mkbhd'), or for_username.
Args: channel_id: YouTube Channel ID (e.g., 'UCBJycsmduvYEL83R_U4JriQ'). for_handle: YouTube Channel handle with or without '@' (e.g., '@mkbhd' or 'veritasium'). for_username: YouTube channel legacy username. raw: If True, returns unaltered raw YouTube Data API response.
Inputs: raw, channel_id, for_handle, for_username. Returns: result
Fetch the latest video uploads from a YouTube channel using the public Atom RSS feed.
Consumes 0 Google Cloud API quota units and requires NO API key. Ideal for monitoring recent uploads without burning quota.
Args: channel_id: The YouTube Channel ID (e.g., 'UCuAXFkgsw1L7xaCfnd5JJOw') or channel URL. max_results: Maximum recent uploads to return (1 to 15, default 15).
Inputs: channel_id (required), max_results. Returns: result
Retrieve video items contained within a YouTube playlist.
Args: playlist_id: The ID of the playlist (e.g. 'PLrAXtmErZgOdP_8GzKt233nyGH51l50DH' or channel uploads ID). max_results: Number of items to retrieve (up to 50, default 20). page_token: Token for retrieving next page of results. raw: If True, returns unaltered raw YouTube Data API response.
Inputs: playlist_id (required), raw, page_token, max_results. Returns: result
Discover real-time trending topics, breakout niches, and viral videos right now.
Analyzes YouTube's official real-time trending chart to extract top rising tags, recurring title keywords, and breakout creators dominating the algorithm.
Args: region_code: Country code (default 'US', e.g. 'GB', 'CA', 'DE', 'FR', 'IN'). category: Niche category ('all', 'tech', 'gaming', 'education', 'howto', 'entertainment', 'news', 'music', 'sports') or numeric ID. max_results: Number of trending videos to analyze (up to 50, default 25). raw: If True, returns unaltered raw YouTube Data API response.
Inputs: raw, category, max_results, region_code. Returns: result
Retrieve top-level comment threads and discussions for a video.
Args: video_id: Video ID or URL to fetch comments for. max_results: Number of comments to retrieve (up to 100, default 20). order: Sort order: 'relevance' (most popular) or 'time' (newest first). page_token: Token for retrieving next page of comments. raw: If True, returns unaltered raw YouTube Data API response.
Inputs: video_id (required), raw, order, page_token, +1 more. Returns: result
Retrieve metadata, views, likes, duration, and tags for one or more YouTube videos.
Args: video_ids: List of 11-character video IDs or URLs (up to 50 IDs). raw: If True, returns unaltered raw YouTube Data API response.
Inputs: video_ids (required), raw. Returns: result
Retrieve full video transcripts or subtitles without requiring an API key.
Args: video_id_or_url: YouTube video ID or full URL (e.g., 'dQw4w9WgXcQ' or 'https://youtu.be/...'). languages: Language priority list (e.g., ['en', 'es']). Defaults to ['en']. format: Output format: 'text' (concatenated string), 'timestamped' (lines with [hh:mm:ss]), or 'json' (array of segment objects). start_seconds: Optional start timestamp to filter segments. end_seconds: Optional end timestamp to filter segments.
Inputs: video_id_or_url (required), format, languages, end_seconds, +1 more. Returns: result
Analyze video script or transcript pacing to predict viewer drop-off points and suggest pattern interrupts.
Calculates words-per-minute (WPM) across segments, flags flat/monotonous monologue stretches (>45 seconds without visual change or question), and injects timestamped retention resets.
Args: script_or_transcript: Raw text of the draft video script or spoken transcript. video_id_or_url: Optional YouTube Video ID or URL to fetch and evaluate live transcript. target_duration_minutes: Optional target video runtime in minutes.
Inputs: video_id_or_url, script_or_transcript, target_duration_minutes. Returns: result
Perform full end-to-end reverse engineering on any YouTube channel.
Deconstructs upload cadence, view-to-sub engagement ratio, title formulas, the opening 60s script hook of their top video, full monetization funnel, and mines real viewer comments for unmet content requests and pain points. Delivers a step-by-step replication playbook for a new creator to model or compete.
Args: channel_id_or_handle: Channel handle (e.g. '@mkbhd', '@aliabdaal'), channel ID ('UC...'), or username. sample_videos: Number of recent uploads to analyze (3 to 15, default 5). include_audience_gaps: If True, mines the comment section of their top video for unmet viewer needs.
Inputs: channel_id_or_handle (required), sample_videos, include_audience_gaps. Returns: result
Launch a multi-niche discovery campaign to identify and rank top YouTube channels.
Useful for creator scouting, influencer marketing campaigns, competitor analysis, and sponsorship outreach.
Args: niches: List of niche keywords/topics (e.g. ['ai automation', 'saas tools', 'b2b sales']). min_subscribers: Minimum subscriber threshold (e.g. 5000 or 10000 for micro-influencers). max_subscribers: Optional maximum subscriber threshold (e.g. 100000). min_videos: Minimum video count to filter out inactive accounts (default: 1). region_code: ISO 3166-1 alpha-2 country code (e.g. 'US', 'GB', 'FR'). channels_per_niche: Number of top channels to return per niche (default: 10). sort_by: Ranking attribute: 'subscribers' (default), 'views', 'videos', or 'avg_views'.
Inputs: niches (required), sort_by, min_videos, region_code, +3 more. Returns: result
Search YouTube specifically for channels matching a query or niche, enriched with subscribers, views, and video counts.
Ideal for creator discovery, niche research, and competitor mapping.
Args: query: Topic or niche keywords (e.g. 'ai automation', 'budget travel', 'investing for beginners'). max_results: Number of channels to return (1 to 50, default 10). order: Ranking order ('relevance', 'videoCount', 'viewCount', 'rating'). region_code: ISO 3166-1 alpha-2 country code (e.g. 'US', 'GB', 'CA').
Inputs: query (required), order, max_results, region_code. Returns: result
Search YouTube for videos, channels, or playlists.
Args: query: Search keywords or query string. max_results: Number of results to return (1 to 50, default 10). search_type: Type of resource to search for ('video', 'channel', 'playlist'). order: Sort order ('relevance', 'date', 'rating', 'viewCount', 'title'). published_after: RFC 3339 formatted date-time string (e.g. '2024-01-01T00:00:00Z'). region_code: ISO 3166-1 alpha-2 country code (e.g. 'US', 'FR'). raw: If True, returns unaltered raw YouTube Data API response.
Inputs: query (required), raw, order, max_results, +3 more. Returns: result
Grade and simulate the click-through-rate (CTR) potential of candidate video titles.
Evaluates titles against YouTube psychological click triggers: curiosity gaps, loss aversion, specificity/numbers, power words, and mobile length sweet-spots (<50 chars). Designates the highest-CTR winning title and provides 3 optimized variations for each candidate.
Args: titles: List of 1 to 8 candidate video titles to test against each other. target_niche: Optional niche context (e.g. 'coding', 'finance', 'gaming').
Inputs: titles (required), target_niche. Returns: result
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 33 tools
Many tools occupy overlapping conceptual territory: audit_channel_strategy and reverse_engineer_channel both deconstruct the same channel metrics, hooks, monetization, and comments, while scout_niche_channels, find_breakout_growth_channels, find_viral_outliers, and find_content_gaps all target channel/niche opportunity discovery. Descriptions help, but the boundaries are blurred enough that an agent could regularly misselect.
All tool names consistently follow a snake_case verb_noun pattern with no mingled conventions. Even where verbs are broad, the naming style is predictable and readable.
33 tools is excessive for a coherent server, especially given the number of overlapping strategy/analysis tools. Many could be consolidated into fewer composite research or optimization tools without losing real capability.
For a creator-research and content-strategy server, the coverage is strong: retrieval, transcripts, comments, channel audits, trend discovery, ideation, SEO metadata, and monetization planning are all represented. It lacks write/management operations like uploading or moderating content, but those appear out of scope for this toolset.
Search YouTube videos, read transcripts and get public video information in your AI assistant.
YouTube for AI agents: transcripts, comments, video stats, channel monitor, Shorts captions.
Fetch YouTube transcripts, monitor channels and track keyword matches from your AI assistant.
YouTube transcripts, search, channel/playlist listings and upload tracking for AI agents.
