The Problem
AI agents using only their training data or a single search source give stale, incomplete answers. A user asking about a product gets outdated prices. A user asking about a topic misses Reddit discussions, YouTube tutorials, and TikTok trends. Single-source agents have blind spots that users notice immediately.
The Scavio Solution
Build a multi-platform retrieval function that queries Google, Amazon, YouTube, Walmart, Reddit, and TikTok through Scavio's unified API. Route each user query to the 2-3 most relevant platforms based on intent detection. Merge results into a single context block for the LLM. One API key, one endpoint, six platforms of fresh data.
Before
Agent queries only Google. Misses Amazon pricing, Reddit opinions, YouTube tutorials, TikTok trends. User asks 'best budget headphones' and gets blog posts from 2024. No price data, no user reviews, no video comparisons.
After
Agent queries Google (articles), Amazon (current prices), Reddit (user opinions), and YouTube (reviews) in parallel. User gets a grounded answer with today's prices, real user feedback, and links to video reviews. Four API calls = $0.02.
Who It Is For
AI agent developers who want their agents to answer questions with fresh, multi-source data instead of relying on stale training knowledge.
Key Benefits
- Six platforms of fresh data through one API key
- Intent-based routing sends queries to the most relevant platforms
- Parallel queries keep latency under 3 seconds
- Four-platform query costs $0.02 total
- Agent answers are verifiable with cited sources from each platform
Python Example
import requests, os
from concurrent.futures import ThreadPoolExecutor
API_KEY = os.environ["SCAVIO_API_KEY"]
H = {"x-api-key": API_KEY, "Content-Type": "application/json"}
PLATFORM_MAP = {
"price": ["amazon", "walmart"],
"review": ["google", "youtube", "reddit"],
"trend": ["tiktok", "youtube", "google"],
"general": ["google", "reddit"],
}
def search_platform(query: str, platform: str) -> dict:
resp = requests.post(
"https://api.scavio.dev/api/v1/search",
headers=H,
json={"query": query, "platform": platform, "country_code": "us"},
timeout=10,
)
return {"platform": platform, "results": resp.json()}
def multi_platform_search(query: str, intent: str = "general") -> list[dict]:
platforms = PLATFORM_MAP.get(intent, PLATFORM_MAP["general"])
with ThreadPoolExecutor(max_workers=4) as pool:
futures = [pool.submit(search_platform, query, p) for p in platforms]
return [f.result() for f in futures]
# Agent retrieval: fresh data from multiple platforms
results = multi_platform_search("best budget headphones 2026", intent="review")
for r in results:
print(f"--- {r['platform']} ---")
for item in r["results"].get("organic_results", [])[:3]:
print(f" {item['title']}")JavaScript Example
const API_KEY = process.env.SCAVIO_API_KEY;
const H = {"x-api-key": API_KEY, "Content-Type": "application/json"};
const PLATFORM_MAP = {
price: ["amazon", "walmart"],
review: ["google", "youtube", "reddit"],
trend: ["tiktok", "youtube", "google"],
general: ["google", "reddit"],
};
async function searchPlatform(query, platform) {
const res = await fetch("https://api.scavio.dev/api/v1/search", {
method: "POST",
headers: H,
body: JSON.stringify({ query, platform, country_code: "us" }),
});
return { platform, results: await res.json() };
}
async function multiPlatformSearch(query, intent = "general") {
const platforms = PLATFORM_MAP[intent] || PLATFORM_MAP.general;
return Promise.all(platforms.map(p => searchPlatform(query, p)));
}
const results = await multiPlatformSearch("best budget headphones 2026", "review");
for (const r of results) {
console.log(`--- ${r.platform} ---`);
for (const item of (r.results.organic_results || []).slice(0, 3)) {
console.log(` ${item.title}`);
}
}Platforms Used
Web search with knowledge graph, PAA, and AI overviews
Amazon
Product search with prices, ratings, and reviews
YouTube
Video search with transcripts and metadata
Walmart
Product search with pricing and fulfillment data
Community, posts & threaded comments from any subreddit
TikTok
Trending video, creator, and product discovery