langchain-scavio

47 LangChain tools for real-time search across Google, Amazon, Walmart, YouTube, Reddit, TikTok, TikTok Shop, and Instagram -- structured data with knowledge graphs, all through a single package.

pip install langchain-scavio

Get your free API key at dashboard.scavio.dev.

Why Scavio over Tavily?

Scavio is a full Tavily alternative built for multi-platform agents — here is Tavily vs Scavio at a glance:

What Can You Build?

• Shopping agents -- search Amazon and Walmart, compare prices, find deals across 22 marketplaces
• Product research agents -- Google reviews + Amazon listings + YouTube reviews + Reddit opinions in one query
• Content research agents -- YouTube trends + Reddit sentiment + Google news in a single workflow
• Brand monitoring -- track what Reddit and Google say about any topic in real time
• Social media agents -- TikTok and Instagram profile analytics, hashtag tracking, post/video comments, and trend discovery

Quick Start

import os
from langchain_scavio import ScavioSearch

os.environ["SCAVIO_API_KEY"] = "sk_live_..."

tool = ScavioSearch()
result = tool.invoke({"query": "best python web frameworks 2026"})

All 47 Tools

Use with a LangChain Agent

Scavio tools plug into the current create_agent API from langchain.agents:

from langchain.agents import create_agent
from langchain_scavio import (
    ScavioSearch,
    ScavioAmazonSearch, ScavioAmazonProduct,
    ScavioWalmartSearch,
    ScavioYouTubeSearch, ScavioYouTubeVideo, ScavioYouTubeTranscript,
    ScavioRedditSearch, ScavioRedditPost,
    ScavioTikTokSearchVideos, ScavioTikTokProfile, ScavioTikTokVideo,
)

agent = create_agent(
    "openai:gpt-5.5",
    tools=[
        ScavioSearch(max_results=5),
        ScavioAmazonSearch(max_results=5),
        ScavioAmazonProduct(),
        ScavioWalmartSearch(max_results=5),
        ScavioYouTubeSearch(max_results=5),
        ScavioYouTubeVideo(),
        ScavioYouTubeTranscript(),
        ScavioRedditSearch(max_results=5),
        ScavioRedditPost(),
        ScavioTikTokSearchVideos(max_results=5),
        ScavioTikTokProfile(),
        ScavioTikTokVideo(),
    ],
)

response = agent.invoke({
    "messages": [{"role": "user", "content": "Find me a Python book on Amazon under $30"}]
})

Async Support

All tools support async invocation:

result = await tool.ainvoke({"query": "async python frameworks"})

Configuration

Google Search

from langchain_scavio import ScavioSearch

tool = ScavioSearch(
    scavio_api_key="sk_live_...",       # or SCAVIO_API_KEY env var
    max_results=5,
    light_request=None,                  # deprecated, ignored (v2 always full, 1 credit)
    include_knowledge_graph=True,
    include_questions=True,
    include_related=False,
    country_code="us",
    language="en",
    search_type="classic",               # classic|news|maps
    device="desktop",
)

Amazon

from langchain_scavio import ScavioAmazonSearch, ScavioAmazonProduct, ScavioAmazonOffers

search = ScavioAmazonSearch(max_results=5)
search.invoke({"query": "wireless headphones", "country": "us", "page": 1})

product = ScavioAmazonProduct()
product.invoke({"query": "B08N5WRWNW"})           # query = ASIN

offers = ScavioAmazonOffers()
offers.invoke({"query": "B08N5WRWNW"})            # every seller for that ASIN

Targeting a marketplace: country takes a two-letter code, not a domain. Supported: us (default), gb (the UK is gb, not uk), ca, de, fr, es, it, jp, in, au, br, mx, nl, pl, se, sg, ae, sa, eg, cn, be, tr. An unrecognised code falls back to us.

Amazon changed in 3.0 (breaking). The upstream provider moved and the request surface shrank. sort_by, pages, category_id, merchant_id, language, currency, device, zip_code and autoselect_variant are gone from all Amazon tools -- the marketplace ignores every one of them, so they are removed rather than kept as silent no-ops (sort_by was verified: all six sort values return the identical unordered set). domain and start_page still work on the wire and are still forwarded, but they are no longer in the tool schemas: use country and page. Response fields were renamed too -- url_image is now image, best_seller/is_amazons_choice collapsed into badge, and buybox is gone (use ScavioAmazonOffers).

Walmart

from langchain_scavio import ScavioWalmartSearch, ScavioWalmartProduct

search = ScavioWalmartSearch(max_results=5)
result = search.invoke({
    "query": "air fryer",
    "sort_by": "price_low",              # best_match|price_low|price_high|best_seller
    "max_price": 5000,                   # in cents
    "fulfillment_speed": "2_days",       # today|tomorrow|2_days|anytime
})

product = ScavioWalmartProduct()
result = product.invoke({"product_id": "123456789"})

YouTube

from langchain_scavio import (
    ScavioYouTubeSearch, ScavioYouTubeVideo, ScavioYouTubeComments,
    ScavioYouTubeTranscript, ScavioYouTubeChannel,
    ScavioYouTubeChannelVideos, ScavioYouTubeStreams,
)

# Video search (2 credits per call)
search = ScavioYouTubeSearch(max_results=5)
result = search.invoke({
    "query": "python tutorial",
    "duration": "medium",                # short|medium|long
    "upload_date": "this_month",         # last_hour|today|this_week|this_month|this_year
    "sort_by": "view_count",             # relevance|date|view_count|rating
    "video_type": "video",               # video|channel|playlist
    "features": ["hd", "subtitles"],     # hd|4k|subtitles|creative_commons|live|360|3d|hdr|vr180
})

# Paginate with the previous response's next_cursor
next_page = search.invoke({
    "query": "python tutorial",
    "cursor": result["data"]["next_cursor"],
})

# Full video details (chapters, captions, keywords)
video = ScavioYouTubeVideo()
result = video.invoke({"video_id": "dQw4w9WgXcQ"})  # video ID or watch URL

# Comments (paginate via data.next_cursor)
comments = ScavioYouTubeComments(max_results=10)
result = comments.invoke({"video_id": "dQw4w9WgXcQ"})

# Transcript as plain text or timed SRT (8 credits per call)
transcript = ScavioYouTubeTranscript()
result = transcript.invoke({"video_id": "dQw4w9WgXcQ", "format": "text"})

# Channel details and uploads
channel = ScavioYouTubeChannel()
result = channel.invoke({"channel_id": "@YouTube"})  # ID, @handle, or URL

channel_videos = ScavioYouTubeChannelVideos(max_results=5)
result = channel_videos.invoke({"channel_id": "UC_x5XG1OV2P6uZZ5FSM9Ttw"})

# Playable/downloadable stream URLs (3 credits per call)
streams = ScavioYouTubeStreams()
result = streams.invoke({"video_id": "dQw4w9WgXcQ"})

# ScavioYouTubeMetadata is a deprecated alias of ScavioYouTubeVideo

Reddit

Reddit endpoints cost 1 credit each.

from langchain_scavio import ScavioRedditSearch, ScavioRedditPost

search = ScavioRedditSearch(max_results=5)
result = search.invoke({"query": "langchain"})
# result["data"]["results"] + next_cursor + has_more
# Relevance order only: the endpoint has no sort or result-type filter

# Paginate by passing back the previous response's next_cursor
next_page = search.invoke({
    "query": "langchain",
    "cursor": result["data"]["next_cursor"],
})

post = ScavioRedditPost()
result = post.invoke({
    "url": "https://www.reddit.com/r/programming/comments/abc123/example_post/"
})
# result["data"] is a flat post object (post_id, title, text, score,
# upvote_ratio, num_comments, media). It does NOT return comments.

TikTok

from langchain_scavio import (
    ScavioTikTokProfile, ScavioTikTokUserPosts, ScavioTikTokVideo,
    ScavioTikTokVideoComments, ScavioTikTokCommentReplies,
    ScavioTikTokSearchVideos, ScavioTikTokSearchUsers,
    ScavioTikTokHashtag, ScavioTikTokHashtagVideos,
    ScavioTikTokUserFollowers, ScavioTikTokUserFollowings,
)

# Look up a user profile (returns sec_uid needed by other tools)
profile = ScavioTikTokProfile()
result = profile.invoke({"username": "tiktok"})
sec_uid = result["data"]["user"]["sec_uid"]

# Fetch their recent posts
posts = ScavioTikTokUserPosts(max_results=5)
result = posts.invoke({"sec_user_id": sec_uid, "sort_type": "1"})  # popular

# Search videos by keyword
search = ScavioTikTokSearchVideos(max_results=5)
result = search.invoke({
    "keyword": "python tutorial",
    "sort_type": "1",                        # 0=relevance, 1=most likes
    "publish_time": "30",                    # 0=all, 1=day, 7=week, 30=month
})

# Get video details and comments
video = ScavioTikTokVideo()
result = video.invoke({"video_id": "7123456789012345678"})

comments = ScavioTikTokVideoComments(max_results=10)
result = comments.invoke({"video_id": "7123456789012345678"})

# Hashtag research
hashtag = ScavioTikTokHashtag()
result = hashtag.invoke({"hashtag_name": "python"})
hashtag_id = result["data"]["challengeInfo"]["challenge"]["id"]

hashtag_videos = ScavioTikTokHashtagVideos(max_results=5)
result = hashtag_videos.invoke({"hashtag_id": hashtag_id})

TikTok Shop

Eight tools over the TikTok Shop catalog. Two things to know before you wire them together:

1. ScavioTikTokShopProduct resolves only about 44% of the product ids that ScavioTikTokShopSearch returns. Upstream has no detail data for the rest, so a not-found result is a normal outcome rather than an error -- skip the product instead of retrying. Search is a listing source, not the first leg of a reliable search-then-detail pipeline.
2. ScavioTikTokShopProduct does not return a price. Upstream masks the digits on the product page, so price.current and price.original come back null. Exact prices are on ScavioTikTokShopSearch, ScavioTikTokShopShopProducts and ScavioTikTokShopCategoryProducts.

from langchain_scavio import (
    ScavioTikTokShopSearch, ScavioTikTokShopSearchSuggestions,
    ScavioTikTokShopProduct, ScavioTikTokShopProductReviews,
    ScavioTikTokShopCategories, ScavioTikTokShopCategoryProducts,
    ScavioTikTokShopShopProducts, ScavioTikTokShopResolve,
)

# Search the US catalog -- this is where exact prices live
search = ScavioTikTokShopSearch(max_results=10)
result = search.invoke({"search": "phone case"})
for product in result["data"]["products"]:
    print(product["title"], product["price"]["current"], product["rating"]["score"])

# Paginate with the opaque cursor; dedupe by product_id across pages
if result["data"]["has_more"]:
    page2 = search.invoke({
        "search": "phone case",
        "cursor": result["data"]["next_cursor"],
    })

# Product detail: rich, but priceless (literally) and only ~44% resolvable
detail = ScavioTikTokShopProduct()
result = detail.invoke({"product_id": "1732293553906094315"})
if result.get("not_found"):
    pass                                     # normal: skip it, do not retry
else:
    result["data"]["variants"]               # stock per SKU
    result["data"]["shop"]["followers_count"]

# Reviews: page with has_more, never with total_reviews (it drifts)
reviews = ScavioTikTokShopProductReviews(max_results=20)
result = reviews.invoke({
    "product_id": "1732293553906094315",
    "page_size": 100,
    "sort": "relevant",                      # "recent" is fresher but text-sparse
    "has_media": True,
})

# Category browse (US and GB only)
categories = ScavioTikTokShopCategories()
tree = categories.invoke({})
category_id = tree["data"]["categories"][0]["category_id"]

listing = ScavioTikTokShopCategoryProducts(max_results=10)
result = listing.invoke({"category_id": category_id})

# A seller's whole catalog, with exact prices
shop = ScavioTikTokShopShopProducts(max_results=10)
result = shop.invoke({"shop_id": "7495514739648989419"})

# Turn any share link into an id
resolve = ScavioTikTokShopResolve()
result = resolve.invoke({"url": "https://vt.tiktok.com/ZT2AHoGsE/"})
result["data"]["product_id"], result["data"]["type"]

# Keyword expansion, the only endpoint with genuine 8-region coverage
suggestions = ScavioTikTokShopSearchSuggestions()
result = suggestions.invoke({"search": "wireless", "region": "GB"})
result["data"]["suggestions"]                # plain strings, no volume or score

Instagram

Instagram is priced per endpoint, not flat: 10 credits by default, 8 for ScavioInstagramPost and ScavioInstagramCommentReplies, and 2 for ScavioInstagramUserPosts.

from langchain_scavio import (
    ScavioInstagramProfile, ScavioInstagramUserPosts, ScavioInstagramUserReels,
    ScavioInstagramTaggedPosts, ScavioInstagramStories,
    ScavioInstagramPost, ScavioInstagramPostComments,
    ScavioInstagramCommentReplies, ScavioInstagramSearchUsers,
    ScavioInstagramSearchHashtags,
    ScavioInstagramUserFollowers, ScavioInstagramUserFollowings,
)

# Look up a user profile (returns user_id usable by other tools)
profile = ScavioInstagramProfile()
result = profile.invoke({"username": "instagram"})
user_id = result["data"]["user"]["id"]

# Fetch their recent posts and reels
posts = ScavioInstagramUserPosts(max_results=5)
result = posts.invoke({"username": "instagram"})

reels = ScavioInstagramUserReels(max_results=5)
result = reels.invoke({"username": "instagram"})

# Get a single post's details and comments
post = ScavioInstagramPost()
result = post.invoke({"shortcode": "C1a2b3c4d5e"})

comments = ScavioInstagramPostComments(max_results=10)
result = comments.invoke({
    "shortcode": "C1a2b3c4d5e",
    "sort_order": "newest",                  # popular (default) or newest
})

# Search users and hashtags
search_users = ScavioInstagramSearchUsers(max_results=5)
result = search_users.invoke({"keyword": "cooking"})

search_hashtags = ScavioInstagramSearchHashtags(max_results=5)
result = search_hashtags.invoke({"keyword": "travel"})

Agent-Controllable Parameters

ScavioSearch

ScavioAmazonSearch

There is no sort, category, merchant or price filter: the marketplace ignores them. Rank results yourself.

ScavioAmazonProduct / ScavioAmazonOffers

ScavioWalmartSearch

ScavioYouTubeSearch

ScavioYouTubeVideo / ScavioYouTubeMetadata

ScavioYouTubeComments

ScavioYouTubeTranscript

ScavioYouTubeChannel

ScavioYouTubeChannelVideos

ScavioYouTubeStreams

ScavioRedditSearch

Results come back in relevance order. There is no sort or result-type parameter: the endpoint accepts only query and cursor.

ScavioRedditPost

ScavioTikTokProfile

ScavioTikTokUserPosts

ScavioTikTokVideo

ScavioTikTokVideoComments

ScavioTikTokCommentReplies

ScavioTikTokSearchVideos

ScavioTikTokSearchUsers

ScavioTikTokHashtag

ScavioTikTokHashtagVideos

ScavioTikTokUserFollowers / ScavioTikTokUserFollowings

ScavioInstagramProfile / ScavioInstagramStories

ScavioInstagramUserPosts / ScavioInstagramUserReels / ScavioInstagramTaggedPosts / ScavioInstagramUserFollowers / ScavioInstagramUserFollowings

ScavioInstagramPost

ScavioInstagramPostComments

ScavioInstagramCommentReplies

ScavioInstagramSearchUsers / ScavioInstagramSearchHashtags

Error Handling

• Empty results raise ToolException with actionable suggestions for the LLM
• API errors return {"error": "message"} without crashing the agent
• handle_tool_error=True ensures LangChain passes errors to the LLM as context

Architecture

ScavioBaseAPIWrapper                      # Auth, headers, sync/async HTTP POST
  +-- ScavioSearchAPIWrapper              # -> /api/v2/google (+ news, maps)
  +-- ScavioAmazonSearchAPIWrapper        # -> /api/v1/amazon/search
  +-- ScavioAmazonProductAPIWrapper       # -> /api/v1/amazon/product
  +-- ScavioAmazonOffersAPIWrapper        # -> /api/v1/amazon/offers
  +-- ScavioWalmartSearchAPIWrapper       # -> /api/v1/walmart/search
  +-- ScavioWalmartProductAPIWrapper      # -> /api/v1/walmart/product
  +-- ScavioYouTubeSearchAPIWrapper       # -> /api/v1/youtube/search
  +-- ScavioYouTubeVideoAPIWrapper        # -> /api/v1/youtube/video
  +-- ScavioYouTubeMetadataAPIWrapper     # -> /api/v1/youtube/video (deprecated alias)
  +-- ScavioYouTubeCommentsAPIWrapper     # -> /api/v1/youtube/comments
  +-- ScavioYouTubeTranscriptAPIWrapper   # -> /api/v1/youtube/transcript
  +-- ScavioYouTubeChannelAPIWrapper      # -> /api/v1/youtube/channel
  +-- ScavioYouTubeChannelVideosAPIWrapper# -> /api/v1/youtube/channel/videos
  +-- ScavioYouTubeStreamsAPIWrapper      # -> /api/v1/youtube/streams
  +-- ScavioRedditSearchAPIWrapper        # -> /api/v1/reddit/search
  +-- ScavioRedditPostAPIWrapper          # -> /api/v1/reddit/post
  +-- ScavioTikTokProfileAPIWrapper       # -> /api/v1/tiktok/profile
  +-- ScavioTikTokUserPostsAPIWrapper     # -> /api/v1/tiktok/user/posts
  +-- ScavioTikTokVideoAPIWrapper         # -> /api/v1/tiktok/video
  +-- ScavioTikTokVideoCommentsAPIWrapper # -> /api/v1/tiktok/video/comments
  +-- ScavioTikTokCommentRepliesAPIWrapper# -> /api/v1/tiktok/video/comments/replies
  +-- ScavioTikTokSearchVideosAPIWrapper  # -> /api/v1/tiktok/search/videos
  +-- ScavioTikTokSearchUsersAPIWrapper   # -> /api/v1/tiktok/search/users
  +-- ScavioTikTokHashtagAPIWrapper       # -> /api/v1/tiktok/hashtag
  +-- ScavioTikTokHashtagVideosAPIWrapper # -> /api/v1/tiktok/hashtag/videos
  +-- ScavioTikTokUserFollowersAPIWrapper # -> /api/v1/tiktok/user/followers
  +-- ScavioTikTokUserFollowingsAPIWrapper# -> /api/v1/tiktok/user/followings
  +-- ScavioInstagramProfileAPIWrapper        # -> /api/v1/instagram/profile
  +-- ScavioInstagramUserPostsAPIWrapper      # -> /api/v1/instagram/user/posts
  +-- ScavioInstagramUserReelsAPIWrapper      # -> /api/v1/instagram/user/reels
  +-- ScavioInstagramTaggedPostsAPIWrapper    # -> /api/v1/instagram/user/tagged
  +-- ScavioInstagramStoriesAPIWrapper        # -> /api/v1/instagram/user/stories
  +-- ScavioInstagramPostAPIWrapper           # -> /api/v1/instagram/post
  +-- ScavioInstagramPostCommentsAPIWrapper   # -> /api/v1/instagram/post/comments
  +-- ScavioInstagramCommentRepliesAPIWrapper # -> /api/v1/instagram/post/comments/replies
  +-- ScavioInstagramSearchUsersAPIWrapper    # -> /api/v1/instagram/search/users
  +-- ScavioInstagramSearchHashtagsAPIWrapper # -> /api/v1/instagram/search/hashtags
  +-- ScavioInstagramUserFollowersAPIWrapper  # -> /api/v1/instagram/user/followers
  +-- ScavioInstagramUserFollowingsAPIWrapper # -> /api/v1/instagram/user/followings

Each tool splits parameters into init-only (developer-controlled, e.g. max_results, domain) and LLM-controllable (passed via args_schema at invocation time, e.g. query, sort_by).

Migrating from Tavily

- from langchain_tavily import TavilySearch
+ from langchain_scavio import ScavioSearch

- tool = TavilySearch(max_results=5)
+ tool = ScavioSearch(max_results=5)

See the full migration guide for parameter mapping and feature comparison, or read more on migrating from Tavily.

License

MIT

About Scavio

Scavio is a unified search API for AI agents — one API key, structured JSON, no scraping or proxies. A real-time Tavily alternative and SerpAPI alternative with data from:

• Google Search API — SERP results, news, images, maps, and knowledge graph
• Amazon Product API and Walmart Product API — product search and details
• YouTube API, TikTok API, and Instagram API — video and social media data
• Reddit API — posts and threaded comments
• TikTok Shop, X (formerly Twitter), and LinkedIn — product listings, tweets, profiles, company pages, and job listings

X and LinkedIn are covered by the Scavio API but do not yet have tools in this package.

Get a free API key and explore the documentation.