TikTok Product Trend Detection Before Amazon

import os, requests, json
from datetime import datetime
from collections import defaultdict

TOKEN = os.environ["SCAVIO_API_KEY"]
AUTH = {"Authorization": f"Bearer {TOKEN}"}
HEADERS = {"x-api-key": TOKEN}
TIKTOK_BASE = "https://api.scavio.dev/api/v1/tiktok"
SEARCH_BASE = "https://api.scavio.dev/api/v1/search"

def detect_product_trends(niches: list) -> list:
    """Find trending products on TikTok. Cost: 1 credit per niche."""
    trending = []
    for niche in niches:
        resp = requests.post(f"{TIKTOK_BASE}/search", headers=AUTH,
            json={"query": f"{niche} must have", "count": 20})
        videos = resp.json().get("videos", [])

        # Extract product mentions from high-performing videos
        for v in videos:
            plays = v.get("stats", {}).get("playCount", 0)
            if plays < 100000:
                continue  # only care about viral content

            trending.append({
                "niche": niche,
                "description": v.get("desc", "")[:100],
                "plays": plays,
                "likes": v.get("stats", {}).get("diggCount", 0),
                "shares": v.get("stats", {}).get("shareCount", 0),
                "author": v.get("author", {}).get("uniqueId", ""),
                "engagement_rate": (
                    v.get("stats", {}).get("diggCount", 0)
                    / max(plays, 1) * 100
                ),
            })

    # Sort by engagement rate (high engagement = strong signal)
    trending.sort(key=lambda x: x["engagement_rate"], reverse=True)
    return trending

niches = ["kitchen gadget", "skincare tool", "desk setup",
          "cleaning hack", "fitness equipment"]
trends = detect_product_trends(niches)
print(f"Found {len(trends)} trending products across {len(niches)} niches")

def cross_reference_shopping(trends: list, top_n: int = 10) -> list:
    """Check if TikTok trends are already on Amazon/Shopping."""
    analyzed = []
    for trend in trends[:top_n]:
        # Extract product keywords from TikTok description
        desc = trend["description"]

        # Search Google Shopping for the product
        shop_resp = requests.post(SEARCH_BASE, headers=HEADERS, json={
            "query": desc[:50],
            "search_type": "shopping",
            "num_results": 5,
        })
        shopping = shop_resp.json().get("shopping_results", [])

        # Search Google for competition level
        web_resp = requests.post(SEARCH_BASE, headers=HEADERS, json={
            "query": desc[:50], "num_results": 5
        })
        organic = web_resp.json().get("organic_results", [])

        trend["shopping_results"] = len(shopping)
        trend["price_range"] = (
            f"{shopping[0].get('price', 'N/A')} - {shopping[-1].get('price', 'N/A')}"
            if shopping else "not found"
        )
        trend["competition_level"] = (
            "high" if len(shopping) > 10
            else "medium" if len(shopping) > 3
            else "low"
        )
        trend["opportunity_score"] = (
            10 if trend["competition_level"] == "low" and trend["plays"] > 500000
            else 7 if trend["competition_level"] == "medium"
            else 3
        )
        analyzed.append(trend)

    return sorted(analyzed, key=lambda x: x["opportunity_score"], reverse=True)

# 10 trends x 2 searches each = 20 credits = $0.10
opportunities = cross_reference_shopping(trends)
for opp in opportunities[:5]:
    print(f"Score {opp['opportunity_score']}/10: {opp['description'][:50]}")
    print(f"  TikTok: {opp['plays']:,} plays | Shopping: {opp['competition_level']}")
    print(f"  Price range: {opp['price_range']}")

def weekly_trend_report(niches: list) -> dict:
    """Full weekly trend detection report."""
    trends = detect_product_trends(niches)
    opportunities = cross_reference_shopping(trends)

    report = {
        "date": datetime.utcnow().strftime("%Y-%m-%d"),
        "niches_scanned": len(niches),
        "trends_found": len(trends),
        "high_opportunity": [
            o for o in opportunities if o["opportunity_score"] >= 7
        ],
        "total_credits": len(niches) + len(opportunities) * 2,
    }
    return report

# Weekly cost: 5 niches + 20 cross-references = 25 credits = $0.125
# Monthly: ~$0.50
report = weekly_trend_report(niches)
print(f"High opportunity products: {len(report['high_opportunity'])}")