Skip to main content
skill CX Codex

youtube-research · manojbajaj95 — Claude Code skill: YouTube research & API

youtube-research — a skill for Claude Code, Codex, Cursor and other agents that searches YouTube, pulls channel and video stats from the YouTube Data API v3, and breaks down transcripts for hooks and retention patterns.

MIT 200 installs ★ 114 updated 2026-09-18

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add manojbajaj95/claude-gtm-plugin --skill youtube-research -a claude-code
Open on GitHub

What youtube-research does

The skill uses Bash, WebSearch, and WebFetch to cover three workflows: pre-production topic research, forensic transcript deconstruction, and raw YouTube Data API v3 queries for search, stats, comments, and channel info.

API mode requires a YOUTUBE_API_KEY from Google Cloud Console with YouTube Data API v3 enabled. The daily quota is 10,000 units — a single search call costs 100 units, most other calls cost 1 unit each. Heavy research sessions can exhaust the quota quickly because of that search cost.

Topic research saves output to ./youtube/episode/[number]_[topic]/research.md and reads project context from strategy/brand.md, about/me.md, and content/ideas.md. Transcript fetching runs through skills/youtube-research/scripts/fetch_transcript.py, which accepts a URL or video ID.

What it helps with

  • search top videos on a topic and retrieve their stats via YouTube Data API v3
  • deconstruct a competitor transcript to extract hooks and retention mechanics
  • find content gaps by mapping what angles are already saturated in a niche
  • fetch channel statistics by handle or pull top comments from a specific video
  • save structured research output for handoff to the next skill in the workflow

Example prompts for the agent

  • Research the topic 'CI/CD automation' on YouTube and save findings to the episode research file
  • Fetch the transcript for VIDEO_ID and extract the hook structure from the first 30 seconds
  • Get channel stats for @HANDLE and list the top 10 videos by view count

How to install youtube-research

1Skills CLI (Vercel)

npx skills add manojbajaj95/claude-gtm-plugin --skill youtube-research -a claude-code

2Manually: put the files in place

~/.claude/skills/youtube-research/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/youtube-research/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about youtube-research

How do I install youtube-research in Claude Code?

Run npx skills add manojbajaj95/claude-gtm-plugin --skill youtube-research -a claude-code. The skill will be available at ~/.claude/skills/youtube-research/SKILL.md.

What does the youtube-research skill do?

It covers three modes: researching a video topic by surveying competitor content, deconstructing transcripts for viral formulas and retention mechanics, and querying the YouTube Data API v3 directly for video stats, channel info, and comments.

What does the 10,000-unit daily quota mean in practice?

Each search request costs 100 units, so you get roughly 100 searches per day before the quota resets. Most other API calls — video details, comments, channel info — cost 1 unit each and rarely become the bottleneck.

Source text

SKILL.md · MIT
---
name: youtube-research
description: Research YouTube topics, analyze competitor videos, deconstruct viral content, and query the YouTube Data API. Use when researching a video topic before planning, analyzing video transcripts for viral patterns, searching competitor channels, or fetching video and channel stats via the YouTube Data API v3.
allowed-tools: Bash, WebSearch, WebFetch
---

# YouTube Research

## Workspace Context

Read bootstrap context before asking questions: `strategy/brand.md` for brand, audience, offer, channels, tools, constraints, and metrics; `about/me.md` for personal voice; `content/ideas.md` and `content/calendar.md` for content planning. Use legacy product-marketing context files only as fallback. Save generated drafts to `content/<platform>/drafts/YYYY-MM-DD_short-topic-slug.md`, and route durable learnings back to `strategy/brand.md`, `about/me.md`, or `content/ideas.md`.

## Operating Contract

This skill is self-contained for its frontmatter scope: use its local instructions, references, scripts, and assets as the playbook; ask only for missing task-specific inputs; hand off to adjacent skills instead of expanding scope; and return an actionable artifact, decision, plan, draft, or diagnostic.



Three modes in one skill:

1. **Topic Research** — competitive landscape, content gaps, strategic insights before planning a video
2. **Video Analysis** — forensic deconstruction of transcripts to extract viral formulas and retention mechanics
3. **API Queries** — direct YouTube Data API v3 access for search, stats, comments, and channel info

---

## When to Use

- Researching a video topic before planning production
- Analyzing a competitor video to extract what makes it work
- Fetching channel stats, video metrics, or comments via the API
- Identifying content gaps and opportunities in a niche

---

## YouTube Data API Setup

### 1. Get an API Key

1. Go to [Google Cloud Console](https://console.cloud.google.com/) → APIs & Services → Library
2. Enable **YouTube Data API v3**
3. Create Credentials → API Key

```bash
export YOUTUBE_API_KEY="your-api-key-here"
```

> **Important:** When piping curl output, wrap the command in `bash -c '...'` to preserve env vars:
> ```bash
> bash -c 'curl -s "https://..." -H "..." | jq .'
> ```

### 2. Key API Commands

**Search Videos:**
```bash
bash -c 'curl -s "https://www.googleapis.com/youtube/v3/search?part=snippet&q=YOUR_QUERY&type=video&maxResults=10&order=viewCount&key=${YOUTUBE_API_KEY}"' | jq '.items[] | {videoId: .id.videoId, title: .snippet.title, channel: .snippet.channelTitle}'
```

**Get Video Details (stats, duration):**
```bash
bash -c 'curl -s "https://www.googleapis.com/youtube/v3/videos?part=snippet,statistics,contentDetails&id=VIDEO_ID&key=${YOUTUBE_API_KEY}"' | jq '.items[0] | {title: .snippet.title, views: .statistics.viewCount, likes: .statistics.likeCount, duration: .contentDetails.duration}'
```

**Get Channel by Handle:**
```bash
bash -c 'curl -s "https://www.googleapis.com/youtube/v3/channels?part=snippet,statistics&forHandle=@HANDLE&key=${YOUTUBE_API_KEY}"' | jq '.items[0] | {id: .id, title: .snippet.title, subscribers: .statistics.subscriberCount, videos: .statistics.videoCount}'
```

**Get Video Comments:**
```bash
bash -c 'curl -s "https://www.googleapis.com/youtube/v3/commentThreads?part=snippet&videoId=VIDEO_ID&maxResults=20&order=relevance&key=${YOUTUBE_API_KEY}"' | jq '.items[] | {author: .snippet.topLevelComment.snippet.authorDisplayName, text: .snippet.topLevelComment.snippet.textDisplay, likes: .snippet.topLevelComment.snippet.likeCount}'
```

**Get Trending Videos:**
```bash
bash -c 'curl -s "https://www.googleapis.com/youtube/v3/videos?part=snippet,statistics&chart=mostPopular&regionCode=US&maxResults=10&key=${YOUTUBE_API_KEY}"' | jq '.items[] | {title: .snippet.title, channel: .snippet.channelTitle, views: .statistics.viewCount}'
```

**Quota:** 10,000 units/day. Search = 100 units. Most others = 1 unit.

See [YouTube Data API docs](https://developers.google.com/youtube/v3) for full reference.

---

## Mode 1: Topic Research

Conduct research before planning a new video. Focus on insights and big levers — not data dumping.

### Workflow

**Step 0: Create research file**

Save all research to: `./youtube/episode/[episode_number]_[topic_short_name]/research.md`

If it already exists, read it and continue from where it left off.

**Step 1: Understand the topic**
- What problem does this video solve?
- Why would someone click on it?
- What makes it relevant now?

**Step 2: Research your own channel**

Use the API to find related videos you've already published. Document:
- Related videos (title, video ID, URL, key metrics)
- What's already been covered and how to differentiate

**Step 3: Competitor research**

Search for 5–8 top videos on the topic. For each:
- Get video details (views, likes, duration)
- Note the title, angle, and what makes it successful
- Synthesize common patterns and approaches

**Step 4: Content gap analysis**

Document:
- **What's saturated** — 3–5 over-covered angles
- **Gaps (Opportunities)** — rated ⭐⭐⭐ high / ⭐⭐ medium / ⭐ low
- **Recommended focus** — specific angle + unique value proposition

Rating criteria:
- ⭐⭐⭐ High: Significant gap, strong demand, clear differentiation
- ⭐⭐ Medium: Moderate gap, some competition, good potential
- ⭐ Low: Minor gap, heavily competed

### Research File Template

```markdown
# [Episode]: [Topic] - Research

## Episode Overview
**Topic**: [Brief description]
**Target Audience**: [Who this is for]
**Goal**: [What viewers will learn/gain]

## YouTube Research
### Your Previous Videos
[Related videos with metrics]

### Top Competing Videos
[5-8 videos: title, channel, views, angle, what works]

### Key Insights
[Patterns and findings synthesized]

## Content Gap Analysis
### What's Already Well-Covered
[List]

### Content Gaps (Opportunities)
[Rated list with ⭐ ratings]

### Recommended Focus
[Specific angle and unique value proposition]

## Production Notes
**Status**: Research Complete
**Created**: [Date]
```

### Parallel Research

If the host environment supports parallel research, split focused tasks such as competitor search, own-channel review, and comment mining. Otherwise, do them sequentially and synthesize findings after each section.

### Pitfalls

- **Data dumping** — Limit to 5–8 competitors, synthesize patterns instead of listing every video
- **Vague gaps** — "Not much content on this" → identify the specific missing angle
- **Long reports** — Focus on insights and big levers

**Next step:** Use `youtube-content` skill to plan the video based on this research.

---

## Mode 2: Video Analysis

Forensic deconstruction of video transcripts to extract viral formulas, hooks, and retention mechanics.

### Getting the Transcript

**Auto-fetch:**
```bash
python skills/youtube-research/scripts/fetch_transcript.py "YOUTUBE_URL_OR_VIDEO_ID"
```

**Manual paste:** YouTube's built-in transcript (click "..." → "Show transcript") or ytscribe.ai.

### Analysis Framework

Approach the transcript like a crime scene — extract everything systematically. See `reference/analysis-framework.md` for the full checklist and templates.

Analyze these 11 dimensions:

1. **Hook Architecture** — Primary hook (first 3–8s), hook type, secondary hooks, fill-in-blank templates
2. **Structural Blueprint** — Content framework (PAS, Story-Lesson-CTA, List-Depth-Summary), beat map, pacing
3. **Retention Mechanics** — Open loops, pattern interrupts, curiosity gaps, payoff points
4. **Emotional Engineering** — Emotional arc, trigger words, identity hooks, Us vs. Them dynamics
5. **Storytelling Elements** — Narrative framework, character positioning, conflict/stakes, specificity
6. **Linguistic Patterns** — Power phrases, sentence rhythm, repetition, conversational triggers
7. **Algorithm Signals** — Watch time optimizers, engagement bait, share/save triggers
8. **CTA Architecture** — Primary CTA, soft CTAs, timing, value exchange
9. **Viral Coefficient** — Shareability score (1–10), comment bait density, crossover potential
10. **Reusable Templates** — Fill-in-blank opening hooks (3 variations), section templates, transition library
11. **Implementation Playbook** — Top 10 steal-this elements, niche adaptation, A/B test suggestions

### Before Analysis, Collect Context

- Your niche/topic
- Your content style (casual, educational, hype, etc.)
- Target platform and video length goal

### Output Format

Structure output with all 11 sections. End with a **Quick Reference Cheatsheet** — one-page summary of all extracted patterns for rapid implementation.

---

## Tools

- **YouTube API**: `bash -c 'curl ...'` with `$YOUTUBE_API_KEY`
- **MCP (if available)**: `mcp__plugin_yt-content-strategist_youtube-analytics__search_videos`, `get_video_details`, `get_channel_details`
- **Web**: `WebSearch` and `WebFetch` for industry trends and context

© manojbajaj95 · MIT · manojbajaj95/claude-gtm-plugin

Similar skills

All skills for Marketing and content →

youtube-research by other authors comes first.

youtube-research bradautomates youtube-research — a skill for Claude Code, Cursor, Codex and other agents that finds high-performing videos via the TubeLab outlier API and produces a report with replicable hook formulas. CX Codex +1 428installs youtube-research PostPlusAI youtube-research — a skill that pulls public YouTube data: video metrics, channel summaries, and comment samples. Handy when you need evidence about content or audience. CX Codex +1 1,100installs youtube-search ZeroPointRepo youtube-search — a Claude Code skill that searches YouTube videos, channels and playlists through TranscriptAPI, then pulls transcripts for what it finds. Kick it off when research needs video sources, not just written ones. CX Codex +1 1,100installs pp-youtube Justin pp-youtube — a skill for Claude Code, Cursor, Codex and other agents that keeps a local SQLite databank of tracked YouTube channels, fed through the YouTube Data API v3. CX Codex +1 423installs search-youtube Samarth Gupta search-youtube finds YouTube videos, pulls transcripts and metadata, downloads audio and walks a channel's uploads through a single CLI the agent calls on its own. CX Codex +1 ★ 276repo stars youtube-api fetcher-sh The youtube-api skill gives an agent YouTube search and data access through fetcher.sh. Use it for channel monitoring, video research, and trend tracking. CX Codex +1 4,700installs youtube-channel-analytics-dashboard reason-machines youtube-channel-analytics-dashboard — a skill for Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI and OpenCode that tracks YouTube channel metrics through a dashboard. CX Codex +1 613installs Arcmira: YouTube Transcript Search arcmira Arcmira: YouTube Transcript Search is an MCP server for Claude Code that searches YouTube and podcast transcripts for quotes, mentions, ads, and speaker appearances. CX Codex +1 ★ 33repo stars

From the same set41

The whole set →
free-tool-strategy free-tool-strategy is a skill for planning a free calculator, generator, or other tool to attract leads and organic traffic, from idea validation through launch planning. CX Codex +1 217installs referral-program The referral-program skill helps you design and optimize referral and affiliate programs: selecting the right type, structuring incentives, identifying share moments, and defining success metrics. CX Codex +1 215installs crm-integration crm-integration is a Claude Code skill with integration patterns for Close CRM, HubSpot, and Salesforce, covering lead sync, deal creation, and webhook handlers. CX Codex +1 211installs ph-content-recycling ph-content-recycling repurposes Product Hunt launch assets into blog posts, documentation, social threads, and emails. Saves drafts directly to your workspace. CX Codex +1 208installs content-strategy-and-planning content-strategy-and-planning organizes content pillars, editorial schedules, and AI search visibility. Generates drafts and buyer-stage messaging directly inside your project files. CX Codex +1 157installs linkedin Audit a LinkedIn profile, review engagement, or draft posts with this skill. It helps turn profile findings and content ideas into actionable next steps. CX Codex +1 145installs marketing-strategy The marketing-strategy skill helps shape B2B SaaS positioning, assess market opportunity, and prepare sales materials. Use it to connect product strategy with customer segments and competitors. CX Codex +1 144installs bootstrap bootstrap — a skill for Claude Code, Codex, Cursor, GitHub Copilot, and Gemini CLI that runs a staged interview and builds the file structure for a new GTM project. CX Codex +1 127installs

Project sponsors

Companies that keep this analytics open