composio-scavio

Scavio real-time search tools for Composio.

Scavio is a single Search API over 31 platforms -- Google, YouTube, Amazon, Walmart, eBay, Target, Home Depot, Reddit, TikTok, TikTok Shop, Instagram, X, LinkedIn, Threads, Kuaishou, Zillow, Redfin, Booking.com, Airbnb, Tripadvisor, Yelp, Indeed, Glassdoor, the Apple App Store, Google Play, SEC EDGAR, Companies House, G2, Capterra, Google Ads Transparency and the Meta Ad Library -- plus extract, which reads any URL as Markdown, plain text or raw HTML, and the website tools that map a site's URLs, crawl a whole site, or extract a list of pages in one job. This package exposes 194 tools covering every live Scavio endpoint as a Composio custom toolkit, so your agents can pull structured, up-to-date results across any Composio-supported framework (OpenAI, Anthropic, LangChain, CrewAI, and more).

New in 0.5.0: map, crawl and batch (opt-in). enable_crawl=True adds SCAVIO_MAP_WEBSITE, SCAVIO_CRAWL_WEBSITE, SCAVIO_CRAWL_STATUS, SCAVIO_BATCH_EXTRACT and SCAVIO_BATCH_STATUS. Off by default, because a crawl bills per page. Requires scavio>=0.23.0. See Map, crawl and batch.

New in 0.4.0: 21 more platforms, and they are opt-in. Registering all 189 tools at once buries the handful an agent actually wants, so the verticals added in this release are off by default. The ten original platforms and extract stay on (104 tools). Turn a vertical on with its flag -- build_scavio_toolkit(enable_zillow=True) -- or pass all=True for everything. Nothing that worked in 0.3.0 changes, except that Walmart grew from 2 tools to 7.

Walmart was rebuilt (breaking). On the rebuilt Walmart endpoints device, delivery_zip and store_id never existed on it and are gone, start_page is replaced by page, and five endpoints beyond search/product are exposed for the first time (REVIEWS, CATEGORY, OFFERS, SELLER, SELLER_PRODUCTS). SCAVIO_WALMART_OFFERS returns the buy-box seller only -- there is no way to page the other sellers -- and SCAVIO_WALMART_SELLER_PRODUCTS returns roughly the first 40 items with no pagination at all.

Google v1 is gone. /api/v1/google was retired on 2026-08-04 and now returns 410. Every Google tool here runs on /api/v2/google* and takes v2 params natively: gl (two-letter country), hl (language), google_domain, device, and start. start is a 0-based result OFFSET, not a page number -- 0 is the first page, 10 the second, 20 the third. The old country_code / language / page arguments are removed rather than silently remapped, because mapping a page number onto start fetches the wrong page. Google responses are also FLAT: there is no data wrapper, and results are organic_results[] with link and snippet.

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.

Reddit search has no filters. /reddit/search accepts only query and cursor; the API strips anything else, so the old type and sort args narrowed nothing and are removed. It returns data.results with next_cursor and has_more. SCAVIO_REDDIT_POST returns the post itself only -- comments are not included, use SCAVIO_REDDIT_POST_COMMENTS for those. The subreddit and user feeds do return data.posts.

Five LinkedIn endpoints are retired. person/contact, company/people, company/jobs, search/people and search/posts were retired upstream and always answer 410 unbilled, so they are deliberately not registered as tools. Use SCAVIO_LINKEDIN_COMPANY (its featured_employees is a 4-6 person sample) instead of company/people, and SCAVIO_LINKEDIN_SEARCH_JOBS with the company name instead of company/jobs.

Install

pip install composio-scavio

Setup

Get a Scavio API key from the Scavio Dashboard (new accounts get 50 one-time signup credits, no credit card; the free plan does not refill monthly). Set it as an environment variable or pass it directly.

export SCAVIO_API_KEY=sk_...
export COMPOSIO_API_KEY=...   # from https://app.composio.dev

Usage

from composio import Composio
from composio_scavio import build_scavio_toolkit

composio = Composio()

# Build the toolkit; expose only the providers you need.
scavio = build_scavio_toolkit(
    api_key="sk_...",        # or rely on SCAVIO_API_KEY
    enable_google=True,
    enable_amazon=True,
    enable_tiktok=False,
    enable_zillow=True,      # opt-in: off unless you ask for it
)

session = composio.create(
    user_id="user_1",
    experimental={"custom_toolkits": [scavio]},
)

result = session.tools.execute(
    "SCAVIO_GOOGLE_SEARCH",
    arguments={"query": "best search API for AI agents", "gl": "us", "hl": "en"},
)
print(result)

SCAVIO_EXTRACT is the agent's "read this page" primitive and is on by default:

page = session.tools.execute(
    "SCAVIO_EXTRACT",
    arguments={"url": "https://example.com/pricing", "format": "markdown"},
)

Pass all=True to register every tool regardless of the individual flags.

Map, crawl and batch

When one page is not enough, enable_crawl=True adds five website tools. SCAVIO_MAP_WEBSITE answers in one call. SCAVIO_CRAWL_WEBSITE and SCAVIO_BATCH_EXTRACT start a background job, wait up to wait_seconds (default 45, max 300) for it to finish, and return the job's status with its first page of results (10 by default) either way. A job that is still queued or running when the wait ends keeps going: read the rest, free, with SCAVIO_CRAWL_STATUS / SCAVIO_BATCH_STATUS and data.id. That keeps an agent turn bounded while a large crawl (up to 6 hours) runs on.

scavio = build_scavio_toolkit(enable_crawl=True)

# A site's URLs from its sitemaps and start page. 1 credit (normal mode).
session.tools.execute("SCAVIO_MAP_WEBSITE", arguments={"url": "https://hono.dev/docs", "limit": 50})

# Same-site pages, breadth-first, within page / depth / path limits.
job = session.tools.execute(
    "SCAVIO_CRAWL_WEBSITE",
    arguments={"url": "https://hono.dev/docs", "max_pages": 25, "include_paths": ["/docs/**"]},
)
# Page on (or pick up a crawl that outlived the wait). Free.
session.tools.execute("SCAVIO_CRAWL_STATUS", arguments={"id": "<data.id>", "cursor": "<next_cursor>"})

# A list of URLs in one job, each extracted exactly like SCAVIO_EXTRACT.
session.tools.execute("SCAVIO_BATCH_EXTRACT", arguments={"urls": ["https://example.com/", "https://hono.dev/docs"]})

Pricing. Map costs 1 credit in normal mode (advanced up to 2, ultra up to 10, priced by what the call needed). Crawl and batch bill per successful page at extract's price for the mode -- normal and advanced 1 credit, ultra 2 -- while failed and skipped pages are free; if your balance runs out the job stops cleanly with stopped_reason: "insufficient_credits". Reading a job is free. Crawl and map obey the site's robots.txt; results are kept for 3 days. Plan limits (pages per crawl / active jobs / URLs per batch): Free 10/1/10, PAYG 100/1/50, Project 500/2/100, Bootstrap 2,000/3/500, Startup 5,000/5/1,000, Growth 10,000/10/1,000. Webhooks, idempotency keys and cancellation are on the REST API and the scavio SDK (client.extract.cancel_crawl(...)), not exposed as agent tools.

Tools

194 tools, all namespaced under the SCAVIO toolkit. Each provider is gated by an enable_* flag. On by default are the ten original platforms and extract (104 tools); the 21 platforms and the 5 website tools marked opt-in are off unless you pass their flag or all=True.

On by default

Opt-in

That is every live Scavio endpoint. Two things are deliberately absent: the deprecated /youtube/metadata alias (identical to SCAVIO_YOUTUBE_VIDEO) and the 5 retired LinkedIn endpoints (always 410, unbilled). Job cancellation is left to the SDK.

Resolve first

Six tools exist only to turn a name you have into the id everything else on that platform is keyed by. Call them first, or the rest of the platform is unreachable: SCAVIO_SEC_LOOKUP (ticker to CIK), SCAVIO_GLASSDOOR_COMPANIES (name to employer_id), SCAVIO_TRIPADVISOR_LOCATIONS (name to geo_id + location_id), SCAVIO_GOOGLE_ADS_ADVERTISERS (brand to advertiser_id), SCAVIO_COMPANIES_HOUSE_SEARCH (name to company_number) and SCAVIO_KUAISHOU_USER_RESOLVE (share link to user_id).

Credits

Every tool states its cost in its own description. Flat-priced platforms:

LinkedIn PERSON, PERSON_ABOUT, COMPANY and POST cost 1 credit.

These surfaces are body-priced -- their cost is a function of the request, not a constant for the route, so no single number can be quoted:

See scavio.dev/docs.

Links

• Scavio: https://scavio.dev
• Docs: https://scavio.dev/docs
• Dashboard: https://dashboard.scavio.dev