Agent skill

tiktok-product-scraper

Step 1 of TikTok workflow. Scrapes product data from tabcut.com OR fastmoss.com using Python scraper. Creates JSON + human-readable MD. Optional video downloads. Designed for n8n orchestration.

View SKILL.md on GitHub Repository
Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/tiktok-product-scraper

SKILL.md

TikTok Product Scraper Skill (Step 1) - Multi-Source Support

WHO RUNS THIS: n8n workflow (or manual terminal) PURPOSE: Scrape product data and optionally download reference videos DATA SOURCES: tabcut.com (default) OR fastmoss.com NEXT STEP: Step 2 - Ad Analysis (gemini-cli)

Overview

This skill scrapes TikTok Shop product data from multiple sources with automatic fallback:

Supported Sources

  1. tabcut.com (default) - Original source with comprehensive data
  2. fastmoss.com (fallback) - Alternative source when Tabcut fails

Auto-Fallback Feature 🆕

When Tabcut returns insufficient data, the scraper automatically retries with FastMoss.

Fallback Triggers (any one = retry with FastMoss):

  • Product name or shop owner is missing/placeholder ("Unknown Product", "undefined", empty, or null)
  • Sales data missing or unparseable (e.g., total_sales or sales_count is null)
  • product_images/ missing or zero images downloaded
  • Top videos list is empty OR all top videos lack a video_url
  • When --download-videos is used: ref_video/ missing or contains 0 MP4s

Example:

bash
python run_scraper.py --product-id 1729724699406473785 --download-videos

# Tabcut attempt...
⚠️ Tabcut data quality check failed:
   - Product name: Unknown Product
   - Total sales: null
   - Images: 0
   - Videos: 0

→ Automatically retrying with FastMoss as fallback source...

✅ FastMoss scraping successful!
   - Product: Gaming Chair
   - Images: 2
   - Videos: 5

Features

  • Product information and sales metrics
  • Top 5 performing videos metadata
  • Optional: Download video files for analysis
  • Always: Convert JSON to human-readable MD for review
  • 🆕 Auto-fallback: Tabcut → FastMoss when data insufficient
  • Choose source: Use --source parameter to force specific source

Prerequisites

✅ Python 3.8+ with virtual environment ✅ Data source credentials in scripts/config/.env

  • Tabcut: TABCUT_USERNAME, TABCUT_PASSWORD
  • FastMoss: FASTMOSS_USERNAME, FASTMOSS_PASSWORD ✅ Playwright browser installed (playwright install chromium) ✅ Dependencies installed (pip install -r requirements.txt)

Setup (one-time):

bash
cd scripts/
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
playwright install chromium

# Configure credentials
cp config/.env.example config/.env
# Edit config/.env with your tabcut.com credentials

Quick Start (n8n)

Option A: Single Product from Tabcut (with videos)

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts
source venv/bin/activate
DATE=YYYYMMDD
OUT="../product_list/$DATE"
python run_scraper.py \
  --product-id {{$json.product_id}} \
  --download-videos \
  --output-dir "$OUT"

# Or explicitly specify tabcut (it's the default)
python run_scraper.py \
  --product-id {{$json.product_id}} \
  --source tabcut \
  --download-videos \
  --output-dir "$OUT"

Option A2: Single Product from FastMoss (with videos)

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts
source venv/bin/activate
DATE=YYYYMMDD
OUT="../product_list/$DATE"
python run_scraper.py \
  --product-id {{$json.product_id}} \
  --source fastmoss \
  --download-videos \
  --output-dir "$OUT"

# Generate human-readable MD
python -c "
import json
import sys
sys.path.append('.')
from tabcut_scraper.utils import json_to_markdown

product_id = '{{$json.product_id}}'
date = 'YYYYMMDD'
with open(f'../product_list/{date}/{product_id}/tabcut_data.json') as f:
    data = json.load(f)
md = json_to_markdown(data)
with open(f'../product_list/{date}/{product_id}/tabcut_data.md', 'w') as f:
    f.write(md)
print('✅ Created tabcut_data.md')
"

Option B: Single Product (no videos)

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts
source venv/bin/activate
DATE=YYYYMMDD
OUT="../product_list/$DATE"
python run_scraper.py --product-id {{$json.product_id}} --output-dir "$OUT"

# Generate MD (same as above)
python -c "..."

Option C: Batch Products from CSV

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts
source venv/bin/activate
DATE=YYYYMMDD
OUT="../product_list/$DATE"
python run_scraper.py \
  --batch-file {{$json.csv_file_path}} \
  --download-videos \
  --output-dir "$OUT"

# Generate MD for all products in batch
python -c "
import json, os, glob
import sys
sys.path.append('.')
from tabcut_scraper.utils import json_to_markdown

date = 'YYYYMMDD'
for json_file in glob.glob(f'../product_list/{date}/*/tabcut_data.json'):
    product_id = json_file.split('/')[-2]
    with open(json_file) as f:
        data = json.load(f)
    md = json_to_markdown(data)
    md_file = json_file.replace('.json', '.md')
    with open(md_file, 'w') as f:
        f.write(md)
    print(f'✅ Created {md_file}')
"

n8n Workflow Node Configuration

Node 1: Execute Command (Scrape)

Command:

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts && \
source venv/bin/activate && \
python run_scraper.py --product-id {{$json.product_id}} {{$json.download_videos && '--download-videos' || ''}} --output-dir ../product_list/{{$json.date}}

Parameters:

  • product_id: String (required) - TikTok product ID
  • download_videos: Boolean (optional, default: false) - Download video files
  • date: String (required) - YYYYMMDD batch folder under product_list/

Expected Output:

json
{
  "success": true,
  "product_id": "1729630936525936882",
  "files_created": [
    "product_list/YYYYMMDD/1729630936525936882/tabcut_data.json"
  ],
  "videos_downloaded": 5
}

Node 2: Convert JSON to MD (Always)

Command:

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688 && \
python3 -c "
import json
import sys

def json_to_markdown(data):
    product_id = data['product_id']
    md = f'''# TikTok Product Data: {product_id}

**Scraped:** {data['scraped_at']}

---

## Product Information

- **Product Name:** {data['product_info']['product_name']}
- **Shop Owner:** {data['product_info']['shop_owner']}
- **Total Sales:** {data['product_info']['total_sales']:,} units
- **Total Revenue:** {data['product_info']['total_sales_revenue']}
- **Product Rating:** {data['product_info'].get('product_rating', 'N/A')}

---

## Sales Data ({data['sales_data']['date_range']})

- **Sales Count:** {data['sales_data']['sales_count']} units
- **Sales Revenue:** {data['sales_data']['sales_revenue']}
- **Related Videos:** {data['sales_data']['related_videos']}
- **Conversion Rate:** {data['sales_data']['conversion_rate']}

---

## Video Analysis Metrics

- **Total Videos:** {data['video_analysis']['带货视频数']}
- **Total Creators:** {data['video_analysis']['带货视频达人数']}
- **Video Sales:** {data['video_analysis']['带货视频销量']}
- **Video Revenue:** {data['video_analysis']['带货视频销售额']}
- **Ad Revenue:** {data['video_analysis']['广告成交金额']}
- **Ad Conversion %:** {data['video_analysis']['广告成交占比']}

---

## Top {len(data['top_videos'])} Performing Videos

'''
    for video in data['top_videos']:
        md += f'''
### Video #{video['rank']}: @{video['creator_username']}

- **Title:** {video['title']}
- **Creator:** @{video['creator_username']} ({video.get('creator_followers', 'N/A')} followers)
- **Published:** {video['publish_date']}
- **Sales:** {video['estimated_sales']} units
- **Revenue:** {video['estimated_revenue']}
- **Views:** {video['total_views']:,}
- **Video URL:** {video.get('video_url', 'N/A')}
- **Local Path:** \`{video.get('local_path', 'Not downloaded')}\`

---
'''
    return md

# Main execution
product_id = '{{$json.product_id}}'
date = '{{$json.date}}'
json_path = f'product_list/{date}/{product_id}/tabcut_data.json'
md_path = f'product_list/{date}/{product_id}/tabcut_data.md'

with open(json_path, 'r') as f:
    data = json.load(f)

md_content = json_to_markdown(data)

with open(md_path, 'w') as f:
    f.write(md_content)

print(json.dumps({
    'success': True,
    'md_file': md_path,
    'product_name': data['product_info']['product_name']
}))
"

Expected Output:

json
{
  "success": true,
  "md_file": "product_list/YYYYMMDD/1729630936525936882/tabcut_data.md",
  "product_name": "HTC NE20 Translator Earbuds..."
}

Manual Execution (Terminal)

Single Product with Videos

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts
source venv/bin/activate
DATE=YYYYMMDD

# Scrape + download videos
python run_scraper.py \
  --product-id 1729630936525936882 \
  --download-videos \
  --output-dir "../product_list/$DATE"

# Convert to MD for review
python3 ../scripts/convert_json_to_md.py 1729630936525936882 --date "$DATE"

Single Product WITHOUT Videos

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts
source venv/bin/activate
DATE=YYYYMMDD

# Scrape metadata only
python run_scraper.py --product-id 1729630936525936882 --output-dir "../product_list/$DATE"

# Convert to MD
python3 ../scripts/convert_json_to_md.py 1729630936525936882 --date "$DATE"

Batch Processing

bash
cd /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts
source venv/bin/activate
DATE=YYYYMMDD

# Create CSV with product IDs
cat > products.csv << EOF
product_id
1729630936525936882
1729479916562717270
1729535919239371775
EOF

# Scrape all (with videos)
python run_scraper.py \
  --batch-file products.csv \
  --download-videos \
  --output-dir "../product_list/$DATE"

# Convert all to MD
for product_dir in ../product_list/$DATE/*/; do
  product_id=$(basename "$product_dir")
  python3 ../scripts/convert_json_to_md.py "$product_id" --date "$DATE"
done

Output Structure

After running this skill:

product_list/YYYYMMDD/{product_id}/
├── tabcut_data.json          # Raw scraped data
├── tabcut_data.md            # ← Human-readable review file (ALWAYS created)
└── ref_video/                # ← Optional (if --download-videos used)
    ├── video_1_{creator}.mp4
    ├── video_2_{creator}.mp4
    ├── video_3_{creator}.mp4
    ├── video_4_{creator}.mp4
    └── video_5_{creator}.mp4

Helper Script: convert_json_to_md.py

Create this utility script:

bash
cat > /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts/convert_json_to_md.py << 'EOF'
#!/usr/bin/env python3
"""Convert tabcut_data.json to human-readable tabcut_data.md"""

import json
import sys
from pathlib import Path

def json_to_markdown(data):
    """Convert scraped JSON data to markdown format"""
    product_id = data['product_id']
    md = f"""# TikTok Product Data: {product_id}

**Scraped:** {data['scraped_at']}

---

## Product Information

- **Product Name:** {data['product_info']['product_name']}
- **Shop Owner:** {data['product_info']['shop_owner']}
- **Total Sales:** {data['product_info']['total_sales']:,} units
- **Total Revenue:** {data['product_info']['total_sales_revenue']}
- **Product Rating:** {data['product_info'].get('product_rating', 'N/A')}
- **Category:** {data['product_info'].get('category', 'N/A')}

---

## Sales Data ({data['sales_data']['date_range']})

- **Sales Count:** {data['sales_data']['sales_count']} units
- **Sales Revenue:** {data['sales_data']['sales_revenue']}
- **Related Videos:** {data['sales_data'].get('related_videos', 'N/A')}
- **Conversion Rate:** {data['sales_data'].get('conversion_rate', 'N/A')}
- **Click-Through Rate:** {data['sales_data'].get('click_through_rate', 'N/A')}

---

## Video Analysis Metrics

- **Total Videos:** {data['video_analysis']['带货视频数']}
- **Total Creators:** {data['video_analysis']['带货视频达人数']}
- **Video Sales:** {data['video_analysis']['带货视频销量']}
- **Video Revenue:** {data['video_analysis']['带货视频销售额']}
- **Ad Revenue:** {data['video_analysis']['广告成交金额']}
- **Ad Conversion %:** {data['video_analysis']['广告成交占比']}

---

## Top {len(data['top_videos'])} Performing Videos

"""

    for video in data['top_videos']:
        md += f"""
### Video #{video['rank']}: @{video['creator_username']}

- **Title:** {video['title']}
- **Creator:** @{video['creator_username']} ({video.get('creator_followers', 'N/A')} followers)
- **Published:** {video['publish_date']}
- **Sales:** {video['estimated_sales']} units
- **Revenue:** {video['estimated_revenue']}
- **Views:** {video['total_views']:,}
- **Video URL:** {video.get('video_url', 'N/A')}
- **Video ID:** {video.get('video_id', 'N/A')}
- **Local Path:** `{video.get('local_path', 'Not downloaded')}`

---
"""

    return md

def main():
    if len(sys.argv) < 2:
        print("Usage: python convert_json_to_md.py <product_id> [--date YYYYMMDD]")
        sys.exit(1)

    product_id = sys.argv[1]
    date = None
    if len(sys.argv) >= 4 and sys.argv[2] == "--date":
        date = sys.argv[3]

    base_dir = Path(__file__).parent.parent / "product_list"
    if date:
        base_dir = base_dir / date
    base_dir = base_dir / product_id
    json_path = base_dir / "tabcut_data.json"
    md_path = base_dir / "tabcut_data.md"

    if not json_path.exists():
        print(f"❌ Error: {json_path} not found")
        sys.exit(1)

    with open(json_path, 'r', encoding='utf-8') as f:
        data = json.load(f)

    md_content = json_to_markdown(data)

    with open(md_path, 'w', encoding='utf-8') as f:
        f.write(md_content)

    print(f"✅ Created: {md_path}")
    print(f"   Product: {data['product_info']['product_name']}")
    print(f"   Sales: {data['product_info']['total_sales']:,} units")
    print(f"   Top Videos: {len(data['top_videos'])}")

if __name__ == "__main__":
    main()
EOF

chmod +x /Users/lxt/Movies/TikTok/WZ/lukas_9688/scripts/convert_json_to_md.py

Configuration Options

Source Selection: NEW!

Option 1: Tabcut (default)

bash
python run_scraper.py --product-id <id>
# OR explicitly
python run_scraper.py --product-id <id> --source tabcut
  • Comprehensive data coverage
  • Well-tested data extraction
  • Original default source

Option 2: FastMoss

bash
python run_scraper.py --product-id <id> --source fastmoss
  • Alternative data source
  • Different authentication method (phone + password with "密码登录")
  • May have different data availability (some features require premium account)
  • Outputs to fastmoss_data.json instead of tabcut_data.json

Tip: You can scrape the same product from both sources to compare data!

Download Videos: YES

When to use:

  • Need to analyze video content with gemini-cli
  • Creating new scripts based on reference videos
  • Deep-dive analysis required

Command:

bash
python run_scraper.py --product-id <id> --download-videos --source <tabcut|fastmoss>

Output: JSON + MD + 5 video files (~50-200 MB)

Download Videos: NO

When to use:

  • Only need product metadata and sales data
  • Quick data gathering for reporting
  • Limited storage/bandwidth

Command:

bash
python run_scraper.py --product-id <id> --source <tabcut|fastmoss>

Output: JSON + MD only (~10 KB)

Error Handling

The scraper handles common issues automatically:

Authentication Failures

  • Automatic retry with exponential backoff
  • Check credentials in scripts/config/.env
  • Run with --headed flag to see browser

Empty 7-Day Data

  • Automatic fallback to 30-day data
  • Logged in output

Product Not Found

  • Retry 2x, then log error
  • Check product ID on tabcut.com

Video Download Failures

  • Logged but doesn't block scraping
  • Continues with remaining videos

Quality Checklist

Before proceeding to Step 2 (Ad Analysis):

  • tabcut_data.json created successfully
  • tabcut_data.md created for human review
  • Review tabcut_data.md - verify product data is correct
  • If videos downloaded: Check ref_video/ folder has MP4 files
  • Top 5 videos metadata looks reasonable
  • Sales data and metrics are populated

Troubleshooting

Issue: Scraper fails to login

Solution:

bash
# Run in headed mode to see browser
cd scripts/
source venv/bin/activate
python run_scraper.py --product-id <id> --headed

Check for:

  • CAPTCHA (requires manual solving)
  • Incorrect credentials in config/.env
  • 2FA enabled on account

Issue: No videos downloaded

Possible causes:

  1. Forgot --download-videos flag
  2. Videos not available on tabcut.com
  3. Download timeout (increase in config/.env)

Solution:

bash
# Retry with longer timeout
python run_scraper.py --product-id <id> --download-videos

Issue: Empty/incomplete data

Solution:

  1. Verify product ID exists on tabcut.com
  2. Check if logged in (run with --headed)
  3. Try 30-day data period if 7-day is empty

Next Step: Ad Analysis (Step 2)

After successful scraping:

  1. Review tabcut_data.md for accuracy
  2. Proceed to .skills/tiktok_ad_analysis.md
  3. Run gemini-cli video analysis (if videos downloaded)
  4. Create video_analysis.md and image_analysis.md

Workflow:

Step 1 (This skill) → tabcut_data.json + tabcut_data.md + videos
                    ↓
Step 2 (gemini-cli) → video_analysis.md + image_analysis.md
                    ↓
Step 3 (Claude Code) → 3 TikTok scripts + Campaign_Summary.md

Version History

v2.0.0 (2025-12-29)

  • NEW: Multi-source support - tabcut.com AND fastmoss.com
  • NEW: --source parameter to choose data source
  • FastMoss authentication with "密码登录" flow
  • Both sources use identical data models for consistency
  • Output files: tabcut_data.json or fastmoss_data.json
  • Comprehensive fastmoss_scraper module (auth, extractors, scraper)

v1.0.0 (2025-12-29)

  • Initial release for n8n orchestration
  • Wrapper around existing Python scraper (tabcut.com only)
  • Always creates human-readable MD file
  • Optional video downloads
  • Batch processing support
  • Error handling documented

Didn't find tool you were looking for?

Be as detailed as possible for better results