deep-profiler

YouTube Credibility Profiler — Terminal CLI

Analyzes YouTube content creators for actionable substance vs. empty buzzwords. Extracts transcripts, evaluates credential depth, and persists results to AWS S3.

---

What It Does

deep-profiler is a terminal-based CLI tool that:

1. Extracts transcripts from any YouTube video URL using yt-dlp
2. Evaluates credibility using the Antigravity SDK (or built-in score_credential evaluator in standalone mode)
3. Persists results to AWS S3 for historical analysis
4. Generates reports showing actionable substance vs. empty buzzwords

Use Cases

• Researching YouTube educators before investing time in their content
• Auditing content creator credibility for partnership decisions
• Building datasets of credible vs. low-quality educational content

---

Tech Stack

Component	Technology	Purpose
CLI	Click / argparse	Terminal interface
Transcript Extraction	yt-dlp	YouTube caption + metadata extraction
Credibility Evaluation	Antigravity SDK	LLM-powered content analysis
Storage	boto3 + AWS S3	Persistent transcript + result storage
Config	python-dotenv	Environment variable management
Testing	pytest + pytest-asyncio	Unit + integration tests
Linting	ruff	Code quality
Typing	mypy	Static type checking
Packaging	Poetry	Dependency management

---

Quick Start

Using Poetry (recommended)

git clone https://github.com/JayV42069/deep-profiler.git
cd deep-profiler

poetry install
poetry run profiler --url "https://www.youtube.com/watch?v=XXX"

Using deepcharts venv

$py = "F:\New AI project\deepchartspro\venv\Scripts\python.exe"
& $py -m pip install -e "F:\New AI project\deep-profiler"
& $py -m deep_profiler --url "https://www.youtube.com/watch?v=XXX"

---

CLI Commands

# Analyze a single video
deep_profiler --url "https://www.youtube.com/watch?v=XXX"

# Verbose debug output
deep_profiler --url "https://www.youtube.com/watch?v=XXX" --verbose

# List cached results
deep_profiler --list

# Export results to JSON
deep_profiler --url "https://www.youtube.com/watch?v=XXX" --export results.json

---

Architecture

deep-profiler/
├── src/deep_profiler/
│   ├── __init__.py
│   ├── __main__.py           # CLI entry point
│   ├── cli.py                # Command-line interface
│   ├── core/
│   │   ├── evaluator.py      # Credibility scoring engine
│   │   ├── transcript.py     # yt-dlp transcript extraction
│   │   └── reporter.py       # Report generation
│   └── scrapers/
│       └── youtube.py        # YouTube-specific scraping
├── aws/
│   └── s3_client.py          # AWS S3 persistence
├── tests/
│   ├── test_evaluator.py
│   ├── test_transcript.py
│   └── test_scrapers.py
├── pyproject.toml
└── README.md

---

How It Works

1. Transcript Extraction

Uses yt-dlp to extract auto-generated captions or manual transcripts from YouTube videos. Falls back to title + description if captions are unavailable.

2. Credibility Evaluation

In standalone mode, runs the built-in score_credential evaluator on extracted transcripts. With the Antigravity SDK, leverages LLM-powered analysis for deeper content assessment.

3. AWS S3 Persistence

Results and transcripts are uploaded to S3 for:

• Historical tracking of content creator credibility
• Dataset building for ML models
• Collaborative research across teams

---

Development

# Install dev dependencies
poetry install --with dev

# Run tests
pytest tests/ -v

# Lint
ruff check src/

# Type check
mypy src/

---

License

MIT — see LICENSE for details.