Scavio search and page-extraction tools for the Vercel AI SDK. Give any AI SDK agent real-time search across Google, YouTube, Reddit, Amazon, Walmart, TikTok, and Instagram -- plus scavio_extract, which reads any URL -- via the Scavio API.

This package exposes a curated subset of the Scavio API, chosen for agent use: 13 tools -- 12 over 7 platforms, plus the core extract endpoint, which belongs to no platform. The full surface is 195 endpoints across 31 platforms, adding X, LinkedIn, TikTok Shop, eBay, Target, Home Depot, Zillow, Redfin, Booking, Airbnb, Tripadvisor, Yelp, Indeed, Glassdoor, the App Store, Google Play, SEC, Companies House, G2, Capterra, Google Ads, Meta Ad Library, Threads, Kuaishou and the Google verticals (Maps, Shopping, News, Flights, Hotels, Trends, AI Mode). To reach all of them, point your agent at the Scavio MCP server at mcp.scavio.dev -- 191 tools over every endpoint, nothing to vendor.

Credits are not uniform: most calls cost 1, YouTube search costs 2 and transcript costs 8, Instagram costs 2 to 10 depending on the endpoint, and scavio_extract is priced by the mode you ask for (1 / 1 / 2), not per call.

Amazon changed (breaking). The upstream provider moved in 2026-07: domain is replaced by country, a two-letter marketplace code (us, gb -- the UK is gb, not uk -- de, jp, ...), and sort_by, pages, category_id, merchant_id, language, currency, device, zip_code and autoselect_variant are gone. The marketplace ignores all of them (sort_by returns the identical unordered set for every value), so they are removed rather than kept as silent no-ops. Rank and filter results yourself.

Install

npm install @scavio/ai-sdk ai zod
export SCAVIO_API_KEY=sk_live_your_key

ai and zod are peer dependencies; scavio (the JS SDK) is bundled.

Usage

import { generateText, stepCountIs } from "ai";
import { openai } from "@ai-sdk/openai";
import { scavioSearch, scavioTools } from "@scavio/ai-sdk";

// A single tool:
const res = await generateText({
  model: openai("gpt-5.5"),
  tools: { scavio_search: scavioSearch({ maxResults: 5 }) },
  stopWhen: stepCountIs(3),
  prompt: "Find the official GitHub repo of the Agno framework",
});

// Or all tools at once:
const res2 = await generateText({
  model: openai("gpt-5.5"),
  tools: scavioTools(),
  stopWhen: stepCountIs(3),
  prompt: "Compare prices for a mechanical keyboard on Amazon and Walmart",
});

Tools

Factory	Tool name	Provider
`scavioExtract`	`scavio_extract`	Read any URL as Markdown / text / HTML, `/api/v1/extract` (1 credit on `normal` and `advanced`, 2 on `ultra`)
`scavioSearch`	`scavio_search`	Google web search, `/api/v2/google` (1 credit)
`scavioYoutubeSearch`	`scavio_youtube_search`	YouTube video search (2 credits)
`scavioYoutubeVideo`	`scavio_youtube_video`	YouTube video details (1 credit)
`scavioYoutubeTranscript`	`scavio_youtube_transcript`	YouTube transcript / subtitles (8 credits)
`scavioYoutubeComments`	`scavio_youtube_comments`	YouTube video comments (1 credit)
`scavioYoutubeChannel`	`scavio_youtube_channel`	YouTube channel details (1 credit)
`scavioRedditSearch`	`scavio_reddit_search`	Reddit posts (1 credit)
`scavioAmazonSearch`	`scavio_amazon_search`	Amazon products (1 credit)
`scavioAmazonOffers`	`scavio_amazon_offers`	Amazon seller offers for one ASIN (1 credit)
`scavioWalmartSearch`	`scavio_walmart_search`	Walmart products (1 credit)
`scavioTiktokSearch`	`scavio_tiktok_search`	TikTok videos (1 credit)
`scavioInstagramSearch`	`scavio_instagram_search`	Instagram users (10 credits)
`scavioTools`	all of the above	bundle for `tools:`

Each factory accepts { apiKey?, maxResults?, ...ScavioConfig }. The key falls back to SCAVIO_API_KEY. Get a key at dashboard.scavio.dev - new accounts get 50 free credits, no credit card.

maxResults trims the result list in place. Google returns a flat body, so it trims organic_results; every other endpoint nests its payload under data, so those trim inside data (data.results, data.products, data.aweme_list, data.users, ...) and the { data, response_time, credits_used, credits_remaining } envelope is preserved. scavioExtract ignores maxResults: it returns page content, not a list, so the page comes back whole.

Reading a page

const page = await generateText({
  model: openai("gpt-5.5"),
  tools: { scavio_extract: scavioExtract() },
  stopWhen: stepCountIs(3),
  prompt: "Summarise the pricing page at https://scavio.dev/pricing",
});

scavio_extract takes url plus an optional format (markdown default, text, html) and mode. mode is the price-bearing param: normal is a plain fetch (1 credit), advanced renders the page in a browser for JS-built sites (1 credit), and ultra routes through residential proxies for the hardest bot walls (2 credits). Only a successful read is billed - a dead link, a bot wall or a timeout costs nothing. See the extract docs.

These tools cover the most common surfaces. The full Scavio API - 195 endpoints across 31 platforms (Google, YouTube, Amazon, Walmart, Reddit, TikTok, TikTok Shop, Instagram, X, LinkedIn, eBay, Target, Home Depot, Zillow, Redfin, Booking, Airbnb, Tripadvisor, Yelp, Indeed, Glassdoor, the App Store, Google Play, SEC, Companies House, G2, Capterra, Google Ads, Meta Ad Library, Threads and Kuaishou), plus extract - is available directly through the scavio SDK or the MCP server.