ScrapeBadger — Web Scraping Platform

ScrapeBadger provides REST APIs for three products. Use the web_fetch tool to call these endpoints.

Authentication

Every request needs the SCRAPEBADGER_API_KEY environment variable. Include it as an HTTP header: X-API-Key: $SCRAPEBADGER_API_KEY

Base URL

https://scrapebadger.com

---

Twitter/X API

Users

Get user profile by username:

GET /v1/twitter/users/{username}/by_username

Get user profile by numeric ID:

GET /v1/twitter/users/{user_id}/by_id

Batch get users by usernames (comma-separated):

GET /v1/twitter/users/batch_by_usernames?usernames=elonmusk,OpenAI

Batch get users by IDs (comma-separated):

GET /v1/twitter/users/batch_by_ids?user_ids=44196397,1230113324

Search users:

GET /v1/twitter/users/search_users?query=AI+agents

Get followers:

GET /v1/twitter/users/{username}/followers

Get following:

GET /v1/twitter/users/{username}/followings

Get mentions:

GET /v1/twitter/users/{username}/mentions

Get subscriptions:

GET /v1/twitter/users/{user_id}/subscriptions

Get user articles:

GET /v1/twitter/users/{user_id}/articles

Tweets

Get tweet by ID:

GET /v1/twitter/tweets/tweet/{tweet_id}

Get multiple tweets by IDs:

GET /v1/twitter/tweets/?tweets=123,456,789

Advanced search:

GET /v1/twitter/tweets/advanced_search?query=web+scraping&query_type=Latest&count=20

Get replies:

GET /v1/twitter/tweets/tweet/{tweet_id}/replies

Get quotes:

GET /v1/twitter/tweets/tweet/{tweet_id}/quotes

Get retweeters:

GET /v1/twitter/tweets/tweet/{tweet_id}/retweeters

Get favoriters (likes):

GET /v1/twitter/tweets/tweet/{tweet_id}/favoriters

Get similar tweets:

GET /v1/twitter/tweets/tweet/{tweet_id}/similar

Get edit history:

GET /v1/twitter/tweets/tweet/{tweet_id}/edit_history

Get community notes:

GET /v1/twitter/tweets/tweet/{tweet_id}/community_notes

Get article:

GET /v1/twitter/tweets/article/{article_id}

Communities

Get community:

GET /v1/twitter/communities/{community_id}

Get community tweets:

GET /v1/twitter/communities/{community_id}/tweets

Search communities:

GET /v1/twitter/communities/search?query=AI

Lists

Get list details:

GET /v1/twitter/lists/{list_id}/detail

Get list tweets:

GET /v1/twitter/lists/{list_id}/tweets

Search tweets in list:

GET /v1/twitter/lists/{list_id}/search_tweets?query=keyword

Trends

Get trending topics:

GET /v1/twitter/trends/

Get trends by location (WOEID):

GET /v1/twitter/trends/place/{woeid}

Geo

Search places:

GET /v1/twitter/geo/search?query=New+York

Get place details:

GET /v1/twitter/geo/places/{place_id}

Spaces

Get Space details:

GET /v1/twitter/spaces/{space_id}

Get broadcast details:

GET /v1/twitter/spaces/broadcast/{broadcast_id}

---

Vinted Marketplace API

Search items:

GET /v1/vinted/search?query=nike+shoes&market=fr&page=1&per_page=20

Optional params: price_from, price_to, brand_ids, order

Get item details:

GET /v1/vinted/items/{item_id}?market=fr

Get user profile:

GET /v1/vinted/users/{user_id}?market=fr

Get user's items:

GET /v1/vinted/users/{user_id}/items?market=fr

Search brands:

GET /v1/vinted/brands?keyword=nike&market=fr

List colors:

GET /v1/vinted/colors?market=fr

List statuses:

GET /v1/vinted/statuses?market=fr

List markets:

GET /v1/vinted/markets

Supported markets: fr, de, uk, us, es, it, nl, be, at, pl, pt, cz, lt, and more.

---

Amazon Marketplace API

The domain parameter is the marketplace TLD (com, co.uk, de, fr, co.jp, …), default com. It also sets the geo-matched residential proxy for correct price/availability.

Search the catalog:

GET /v1/amazon/search?query=mechanical+keyboard&domain=com&page=1

Optional params: sort_by, category, min_price, max_price, zip

Keyword autocomplete:

GET /v1/amazon/autocomplete?query=mechan&domain=com

Get product detail (PDP) by ASIN:

GET /v1/amazon/products/{asin}?domain=com

Optional params: zip, language. Returns variants, buybox, badges, A+ content, frequently-bought-together, also-bought.

Get all seller offers + buybox:

GET /v1/amazon/products/{asin}/offers?domain=com

Get product reviews (featured public subset):

GET /v1/amazon/products/{asin}/reviews?domain=com&page=1

Optional params: sort_by (helpful|recent), star, verified_only, media_only

Best Sellers:

GET /v1/amazon/bestsellers?domain=com&category={node}&page=1

New Releases:

GET /v1/amazon/new-releases?domain=com&category={node}&page=1

Deals (Today's / Lightning):

GET /v1/amazon/deals?domain=com&category=...&page=1

Browse a category by node:

GET /v1/amazon/category?node={node}&domain=com&page=1

Optional param: sort_by

Seller profile:

GET /v1/amazon/sellers/{seller_id}?domain=com

Seller storefront products:

GET /v1/amazon/sellers/{seller_id}/products?domain=com&page=1

Seller buyer feedback:

GET /v1/amazon/sellers/{seller_id}/feedback?domain=com&page=1

List markets (free):

GET /v1/amazon/markets

List categories (free):

GET /v1/amazon/categories

Supported marketplaces (20): com, co.uk, de, fr, it, es, nl, se, pl, ca, com.mx, com.br, co.jp, in, com.au, sg, ae, sa, com.tr, eg.

---

eBay Marketplace API

The domain parameter is the marketplace TLD (com, co.uk, de, fr, …), default com. It also sets the geo-matched residential proxy for correct price/availability. Aliases: us→com, uk/gb→co.uk, au→com.au.

Search active listings:

GET /v1/ebay/search?query=mechanical+keyboard&domain=com&page=1

Optional params: category_id, per_page (60|120|240), sort_by (best_match|ending_soonest|newly_listed|price_low_to_high|price_high_to_low), condition (new|open_box|refurbished|used|for_parts|graded|ungraded), buying_format (auction|buy_it_now|best_offer), min_price, max_price, free_shipping, location (domestic|worldwide). Returns results[] + facets + pagination.

Search active listings by image (eBay's own visual search — 10 credits):

POST /v1/ebay/search/by-image
{"image_url": "https://example.com/sneaker.jpg", "domain": "com"}

Send exactly one of image_url (a public http(s) URL we download) or image_base64 (the image itself, JPEG or PNG, max 10 MB decoded; a data:image/jpeg;base64,… URL also works). Same optional filters as /search — category_id, page, per_page, condition, buying_format, min_price, max_price, free_shipping, location, language — and the same response shape, with query: null. There is no sort_by: eBay ignores sorting on a visual results page. An image eBay cannot read is a 400, never an empty result set.

Search completed/sold listings (price research):

GET /v1/ebay/completed?query=iphone+13+pro&domain=com&page=1

Optional params: category_id, per_page, sort_by, condition (incl. graded/ungraded for cards), min_price, max_price, location (domestic keeps only this marketplace's own country — foreign sales otherwise skew a national price series, and eBay shows their prices converted). Returns sold listings (sold=true).

Keyword autocomplete:

GET /v1/ebay/autocomplete?query=mechan&domain=com

Get item detail by item ID:

GET /v1/ebay/items/{item_id}?domain=com

Returns specifics, images, shipping, returns, and seller.

Get item reviews:

GET /v1/ebay/items/{item_id}/reviews?domain=com&page=1

Returns reviews[] + aggregate rating + histogram.

Seller profile:

GET /v1/ebay/sellers/{username}?domain=com

Seller active listings:

GET /v1/ebay/sellers/{username}/items?domain=com&page=1

Optional params: query, per_page.

Seller buyer feedback:

GET /v1/ebay/sellers/{username}/feedback?domain=com&page=1

Browse a category by ID:

GET /v1/ebay/categories/{category_id}/items?domain=com&page=1

Optional params: per_page, sort_by, min_price, max_price.

List categories (free):

GET /v1/ebay/categories

List markets (free):

GET /v1/ebay/markets

Supported marketplaces (18): com, co.uk, de, fr, it, es, com.au, ca, at, ch, be, ie, nl, pl, com.hk, com.sg, com.my, ph.

---

Realtor Real-Estate API

Unified real-estate API over realtor.com (US, USD) and realtor.ca (Canada, CAD). The market parameter (us or ca, default us) picks the portal and the geo-matched residential proxy. Rich datapoints: price + price/tax history, beds/baths/sqft/lot, coordinates, photos/videos/virtual tours, features/amenities, HOA, schools, open houses, agent/office/broker contacts, days-on-market, price-per-sqft.

Search listings:

GET /v1/realtor/search?location=Austin,+TX&market=us&status=for_sale

Required: location (free-text "Austin, TX" / ZIP / "Toronto, ON"; optional if a CA bbox is given). Optional params: status (for_sale|for_rent|sold|pending, default for_sale), price_min, price_max, beds_min, baths_min, sqft_min/sqft_max (US), property_type (US CSV: single_family,condos,townhomes,multi_family,mobile,land), sort (relevant|newest|price_low|price_high|photo_count), page, limit (1-200), lat_min/lat_max/lng_min/lng_max (CA bbox; realtor.ca caps ~600/box). Returns results[] of Property + total/count/page/total_pages.

Get property detail by ID:

GET /v1/realtor/properties/{property_id}?market=us

Returns a full PropertyDetail: price/tax history, schools, amenities, estimates, agents, open houses.

Location autocomplete:

GET /v1/realtor/autocomplete?query=austi&market=us

Optional param: limit (1-25, default 10). Returns suggestions[] (cities, ZIPs, neighborhoods, addresses).

List markets (free):

GET /v1/realtor/markets

Supported markets (2): us (realtor.com), ca (realtor.ca).

Walmart API

Product, seller and store data from walmart.com (US only, USD, en-US). No Walmart account or developer key is needed. walmart.ca and walmart.com.mx run on different platforms and are not supported.

Search products:

GET /v1/walmart/search?query=laptop&sort=price_low&min_price=100&max_price=500

Required: query. Optional: page (1-10), sort (best_match|best_seller|price_low|price_high|rating_high|new), min_price, max_price, facet (Walmart's own syntax, e.g. brand:HP). Returns 40-60 organic products a page.

Browse a category / deals:

GET /v1/walmart/category?path=electronics/3944&page=2
GET /v1/walmart/deals

path comes from a walmart.com/browse/... or /cp/... URL. /category takes no sort — Walmart's browse pages silently ignore it, so it is not offered. /deals is the rollback/clearance feed; no other Walmart API exposes one.

Product detail and reviews:

GET /v1/walmart/products/5689919121
GET /v1/walmart/products/5689919121/reviews?page=1&sort=helpful

item_id is the numeric segment at the end of a /ip/... URL. The product carries UPC/GTIN, structured specifications, variants, per-method promised delivery dates, return policy, and a 10-review sample. The reviews endpoint adds the full star histogram (one_star..five_star counts and percentages) plus Walmart's own most-helpful positive and negative reviews. 10 reviews a page — Walmart's page size, not adjustable.

Marketplace sellers:

GET /v1/walmart/sellers/101040442
GET /v1/walmart/sellers/6907/products?query=camera

The seller id is the numeric catalog_seller_id found on a product — NOT the 32-character hex seller_id, which 404s. The profile endpoint has no product list because Walmart renders a storefront's grid client-side; /products goes through search with a retailer_id facet instead, which is why it requires a query.

Stores:

GET /v1/walmart/stores/100

Returns the store's address, geo, phone, weekly hours, and per-department services (pharmacy, deli, auto centre — each with its own phone and hours), plus ~30 nearby stores. Walmart has no working store-finder-by-ZIP page, so walk outward from a known store id.

Autocomplete:

GET /v1/walmart/autocomplete?query=lapt

Gotchas that will bite you:

• total_results_reported is Walmart's own claim (~14,000) and is not reachable — results dry up after page 10 (search) or 11 (browse), roughly 500 products a query. Use has_more_pages. To go deeper, segment by brand facet or price band rather than paging.
• Consecutive pages overlap ~20% because Walmart re-ranks between requests — deduplicate on us_item_id.
• Facets can be applied but not enumerated; Walmart renders the filter rail client-side.
• Searching by UPC does not work — Walmart returns zero results for a bare UPC, including a product's own.
• Prices are store-specific and Walmart resolves the store from the request IP. It cannot be pinned by a parameter, so every product returns a location block naming the store the prices actually came from.
• HTTP 422 is an anti-bot challenge: not billed, and retrying shortly succeeds.

Cost: 1 credit autocomplete; 5 credits search / category / deals / seller / seller products / store; 10 credits product detail and reviews. markets and health are free.

DuckDuckGo API

Privacy-first search from duckduckgo.com — web, images, news and videos, plus autocomplete and Instant Answers. No DuckDuckGo account or API key is needed. Region is set with the kl/l code (us-en, uk-en, de-de, … wt-wt = all regions); safesearch is on | moderate | off.

Web search:

GET /v1/duckduckgo/search?query=privacy+tools&region=us-en&safesearch=moderate&timelimit=week&page=1

Required: query. Optional: region (default wt-wt), safesearch (default moderate), timelimit (day|week|month|year), page (1-25). Returns organic results (title, url, snippet, display_url, is_ad), the zero-click abstract box, and has_next.

Image search:

GET /v1/duckduckgo/images?query=northern+lights&size=Large&color=Blue&image_type=photo&layout=Wide&license=Public

Required: query. Optional: region, safesearch, page (1-10, 100 results a page), size (Small|Medium|Large|Wallpaper), color, image_type (photo|clipart|gif|transparent|line), layout (Square|Tall|Wide), license (Any|Public|Share|ShareCommercially|Modify).

News search:

GET /v1/duckduckgo/news?query=elections&region=us-en&timelimit=day&page=1

Required: query. Optional: region, safesearch, timelimit (day|week|month|year), page (1-10, 30 results a page). Each item carries date_utc (unix) and date_at (ISO), source, excerpt and image.

Video search:

GET /v1/duckduckgo/videos?query=guitar+lesson&duration=medium&resolution=high&page=1

Required: query. Optional: region, safesearch, page (1-10, 60 results a page), duration (short|medium|long), resolution (high|standard). Returns title, publisher, uploader, duration, views and thumbnails.

Autocomplete:

GET /v1/duckduckgo/autocomplete?query=pyth&region=us-en

Required: query. Optional: region. Returns { query, suggestions[] }.

Instant Answer:

GET /v1/duckduckgo/instant?query=valley+forge

Required: query. DuckDuckGo's Instant Answer API — abstract, definition, direct answer and related topics.

Regions:

GET /v1/duckduckgo/regions

No params. Returns the full list of 64 DuckDuckGo kl region codes (us-en, uk-en, de-de, … wt-wt = all).

Cost: 1 credit autocomplete / instant; 5 credits search / images / news / videos. regions and health are free.

Apartments.com Rentals API

US rental listings from apartments.com (US, USD, en-US) with unit-level pricing — not just a property's advertised range, but every individual rentable unit with its own rent, beds, baths, square footage and availability date. No apartments.com account is needed.

Search listings:

GET /v1/apartments/search?location=kansas-city-mo&beds=1&max_price=1500

Required: location — the apartments.com slug (kansas-city-mo, new-york-ny) or a ZIP (64108). Optional: page (1-28, 40 cards a page), beds (0=studio, 1-4), min_price, max_price. Returns results[] of cards plus total_results / total_pages.

Filters verified against live counts on kansas-city-mo (700 unfiltered): beds=1 → 374, beds=2 → 490, beds=0 → 145, max_price=1500 → 646, beds=1&max_price=1500 → 320. apartments.com also renders a pet filter, but it returns the unfiltered total, so it is deliberately not exposed.

Get property detail (unit-level):

GET /v1/apartments/property?url=https%3A%2F%2Fwww.apartments.com%2Furbane-kansas-city-mo%2Fwcd6e5k%2F
GET /v1/apartments/properties/{slug}/{property_id}

Returns the property (name, split address, geo, phone, amenities, office hours, photos), its floor_plans[], and a flattened units[] across all plans.

IMPORTANT — use rent, not max_term_rent. Every unit carries both. rent is the advertised monthly price the site shows a renter. max_term_rent is apartments.com's raw data-maxrent attribute, which measures roughly twice the advertised rent and appears to be an upper bound across lease terms. It is exposed unparsed for completeness only.

available_text is verbatim ("Now", "Sep 3"); the site renders no year, so it is deliberately not a date. A floor plan with units_available: 0 and valid beds/baths/rent/sqft is the site's own layout for properties that advertise plans rather than individual units — not missing data.

A search card is a summary with a rent/bed rollup; pass its url to the property endpoint for units.

Do not use render_js for apartments.com via the general Web Scraping API — the page is fully server-rendered, so JS rendering only forces the slow browser path (100+ s) and can exceed the CDN timeout with a 524.

Cost: 5 credits per search page and per property.

Idealista Real Estate API

Real-estate listings from Idealista across 3 markets — Spain (es), Italy (it), Portugal (pt), all EUR. The market parameter defaults to es. No upstream Idealista key is needed; ScrapeBadger signs requests to Idealista's mobile API server-side. Search is keyed by location codes, not free text, so use the two-step flow: suggest → search → property detail.

Suggest locations (free-text → location codes):

GET /v1/idealista/suggest?query=sagrada+familia&market=es

Optional params: operation (sale|rent, default sale), property_type (homes|offices|premises|garages|bedrooms|newDevelopments|lands|storageRooms|buildings, default homes). Returns locations[] + total + market. Only entries with type == "location" carry a usable locationId.

Search listings (by location code):

GET /v1/idealista/search?location=0-EU-ES-28-07-001-079&operation=sale&market=es

Optional params: property_type (as above, excluding bedrooms), page (default 1), max_items (default 30, 1..50), sort_by (weigh|price|distance|publicationDate|modificationDate|size|floor|rooms|ratioeurm2|pricedown), sort_order (asc|desc, default desc), min_price, max_price, min_size, max_size, min_rooms, max_rooms. Returns elementList[] + total + totalPages + actualPage + itemsPerPage + summary[] + market.

Full property detail (by property code):

GET /v1/idealista/properties/{property_code}?market=es&locale=en

Optional params: locale (en|es|it|pt, default en). Returns property + market.

List markets (free):

GET /v1/idealista/markets

Supported markets (3): es (idealista.com), it (idealista.it), pt (idealista.pt).

---

Immobiliare Real Estate API

Real-estate listings from the Immobiliare Group — Italy's #1 portal immobiliare.it plus its sister portals — across 4 markets: it (www.immobiliare.it), es (www.indomio.es), gr (www.indomio.gr), lu (www.immotop.lu), all EUR. The market parameter defaults to it. No upstream API key needed. Search accepts a free-text location (auto-resolved) or explicit region_id/province_id/city_id codes from /autocomplete.

Location autocomplete (free-text → location IDs):

GET /v1/immobiliare/autocomplete?query=milano&market=it

Returns suggestions[] — each {id, label, type (nation|region|province|comune|zone), region_id, province_id, city_id, macrozone_ids[], url}.

Search listings:

GET /v1/immobiliare/search?location=Milano&contract=sale&market=it

Optional params: region_id/province_id/city_id (instead of free-text location), contract (sale|rent), category (residential|commercial|garages|offices|land|buildings|warehouses), price_min, price_max, surface_min, surface_max, rooms_min, rooms_max, bathrooms_min, sort (relevance|price_asc|price_desc|newest|oldest|surface_desc|surface_asc), page. Returns listings[] + count + total_ads + current_page + max_pages + is_results_limit_reached + related_searches[].

Get full listing detail:

GET /v1/immobiliare/listings/{id}?market=it

Returns a full Listing: price + €/m², surface, rooms, baths, floor, energy class, GPS coordinates, all photos, virtual tours, agency + agent contacts, multi-unit properties[].

Agency profile:

GET /v1/immobiliare/agencies/{id}?market=it

Agency listings:

GET /v1/immobiliare/agencies/{id}/listings?market=it&page=1

Market price insights (€/m² monthly time series):

GET /v1/immobiliare/market-insights/prices?region_id=10&market=it

Required: region_id. Optional params: province_id, city_id, contract (sale|rent). Returns {contract, unit: "EUR/m²", points[{label, value}]}.

List markets (free):

GET /v1/immobiliare/markets

Reference data — contracts/categories/sorts (free):

GET /v1/immobiliare/reference

Supported markets (4): it (immobiliare.it), es (indomio.es), gr (indomio.gr), lu (immotop.lu).

---

YouTube API

No YouTube Data API key, OAuth, or quota required. Data is fetched via YouTube's unauthenticated InnerTube JSON API (youtubei/v1/*) plus SSR-HTML fallback. The gl parameter (ISO 3166-1 alpha-2 content region, default US) and hl parameter (BCP-47 UI language, default en) steer the geo-matched residential proxy + locale. All list endpoints paginate via a continuation token (no page numbers); pass the returned token back as continuation. All datetime fields ship as both *_utc (Unix) and *_at (ISO 8601 UTC).

Search videos / channels / playlists:

GET /v1/youtube/search?query=machine+learning&gl=US&hl=en

Optional params: type (video|channel|playlist|movie|all), sort_by (relevance|date|views|rating), upload_date (hour|today|week|month|year), duration (short|medium|long), features (csv: hd,4k,360,vr180,3d,hdr,cc,subtitles,live,creative_commons,location,purchased), continuation. Returns results[] + continuation.

Search autocomplete:

GET /v1/youtube/autocomplete?query=machine+lea&gl=US&hl=en

Get video detail (merged player+next):

GET /v1/youtube/videos/{video_id}

Related videos:

GET /v1/youtube/videos/{video_id}/related?continuation=...

Video comments and nested replies:

GET /v1/youtube/videos/{video_id}/comments?sort_by=top&continuation=...
GET /v1/youtube/videos/{video_id}/comments/{comment_id}/replies?continuation=...

Transcript (timed text + full text + SRT/VTT):

GET /v1/youtube/videos/{video_id}/transcript?lang=en

Captions / streams / live chat:

GET /v1/youtube/videos/{video_id}/captions
GET /v1/youtube/videos/{video_id}/streams
GET /v1/youtube/videos/{video_id}/live_chat?continuation=...

Batch video lookup (POST):

POST /v1/youtube/videos/batch
{ "video_ids": ["dQw4w9WgXcQ", "..."] }

Channel detail + tabs:

GET /v1/youtube/channels/{channel_id}
GET /v1/youtube/channels/{channel_id}/videos?sort_by=newest&continuation=...
GET /v1/youtube/channels/{channel_id}/shorts?continuation=...
GET /v1/youtube/channels/{channel_id}/streams?continuation=...
GET /v1/youtube/channels/{channel_id}/playlists?continuation=...
GET /v1/youtube/channels/{channel_id}/community?continuation=...
GET /v1/youtube/channels/{channel_id}/about
GET /v1/youtube/channels/{channel_id}/subscriber_count
GET /v1/youtube/channels/{channel_id}/search?query=...&continuation=...

Resolve a handle / custom URL to a channel_id:

GET /v1/youtube/channels/resolve?handle=@GoogleDevelopers

Playlists and mixes:

GET /v1/youtube/playlists/{playlist_id}
GET /v1/youtube/playlists/{playlist_id}/items?continuation=...
GET /v1/youtube/mixes/{mix_id}

Trending, hashtags, home feed:

GET /v1/youtube/trending?gl=US&type=now&continuation=...
GET /v1/youtube/trending/shorts?gl=US&continuation=...
GET /v1/youtube/hashtags/{tag}?gl=US&continuation=...
GET /v1/youtube/home?gl=US&hl=en&continuation=...

Shorts:

GET /v1/youtube/shorts/{short_id}
GET /v1/youtube/shorts/by_sound/{sound_id}?continuation=...

Community posts:

GET /v1/youtube/posts/{post_id}
GET /v1/youtube/posts/{post_id}/comments?continuation=...

YouTube Music search:

GET /v1/youtube/music/search?query=...

Reference (free):

GET /v1/youtube/oembed?url=...
GET /v1/youtube/categories
GET /v1/youtube/languages
GET /v1/youtube/regions
GET /v1/youtube/markets

---

TikTok API

No TikTok login required. The region parameter (ISO 3166-1 alpha-2, default US) steers the residential proxy geo + locale. count is 1–50. Most endpoints use a fast device-signed mobile path with deep cursor pagination. All datetime fields ship as both *_utc (Unix) and *_at (ISO 8601 UTC).

Users

Get user profile:

GET /v1/tiktok/users/{username}

Get user's videos (cursor pagination):

GET /v1/tiktok/users/{username}/videos?count=30&cursor=...

Get followers (403 if the account hides the list):

GET /v1/tiktok/users/{username}/followers?count=20

Get following (403 if private):

GET /v1/tiktok/users/{username}/following?count=20

Get liked videos (403 if private):

GET /v1/tiktok/users/{username}/liked?count=20

Get reposted videos:

GET /v1/tiktok/users/{username}/reposts?count=20

Videos

Get video / photo detail:

GET /v1/tiktok/videos/{video_id}

Optional param: username

Get video comments:

GET /v1/tiktok/videos/{video_id}/comments?count=20

Get comment replies:

GET /v1/tiktok/comments/{comment_id}/replies?item_id={video_id}&count=20

Get related videos:

GET /v1/tiktok/videos/{video_id}/related?count=20

Get auto-caption transcript:

GET /v1/tiktok/videos/{video_id}/transcript

Get oEmbed metadata for a TikTok URL:

GET /v1/tiktok/oembed?url=https://www.tiktok.com/@user/video/123

Hashtags & Music

Get hashtag / challenge detail:

GET /v1/tiktok/hashtags/{name}

Get videos for a hashtag:

GET /v1/tiktok/hashtags/{name}/videos?count=20

Get sound / music detail:

GET /v1/tiktok/music/{music_id}

Get videos using a sound:

GET /v1/tiktok/music/{music_id}/videos?count=20

Search

General search (returns videos):

GET /v1/tiktok/search?query=ai+agents&count=20

Search videos:

GET /v1/tiktok/search/videos?query=ai+agents&count=20

Search hashtags:

GET /v1/tiktok/search/hashtags?query=ai&count=20

Search users:

GET /v1/tiktok/search/users?query=charli&count=20

Trending

Trending videos (curated Explore feed):

GET /v1/tiktok/trending/videos?region=US&count=20

Trending hashtags (with view_count + creator counts):

GET /v1/tiktok/trending/hashtags?region=US&count=20

Trending songs (ranked by usage):

GET /v1/tiktok/trending/songs?region=US&count=20

Ad Library (EU only)

Search the TikTok Commercial Content Library (EU-DSA ad transparency):

GET /v1/tiktok/ads/search?query=sneakers&region=DE&days=30&count=20

Search by keyword (query) OR by advertiser (advertiser_id). Paginates via offset + search_id. EU regions only.

Reference (free)

List supported regions:

GET /v1/tiktok/regions

Coverage notes: following/liked return 403 when the account hides those lists (most large accounts do). Trending creators and trending keywords are not available (TikTok deprecated the Creative Center trend lists). The Ad Library is EU-only.

---

ChatGPT API

Sends a prompt to the real chatgpt.com (not the OpenAI API) and returns the answer as structured JSON, including the web sources ChatGPT cited. Anonymous — no OpenAI account or API key is involved. Measured latency 20-25s ungrounded, 30-70s with web search per call, so set a generous client timeout and don't poll.

Ask ChatGPT (20 credits):

GET /v1/chatgpt/ask?prompt=best+open-source+vector+databases&country=US&web_search=auto

Required: prompt (max 4096 chars). Optional: country (ISO-3166 alpha-2 egress country, default US), web_search (auto | force | off, default auto). Also accepts POST with the same fields as a JSON body.

Returns: prompt, answer, answer_markdown, citations[], search_results[], source_domains[], web_search_triggered (whether ChatGPT actually browsed), reference_tokens[], model, conversation_id, message_id, country, answer_length, citation_count, latency_ms, created_utc/created_at.

Each Citation is {url, title, snippet, domain, attribution, pub_date_utc, published_at, start_index, end_index, matched_text} — start_index/end_index are char offsets into answer, so you can anchor a source to the exact sentence it supports. search_results[] is the FULL retrieved set ({url, title, snippet, domain, attribution, pub_date_utc, published_at, ref_index, cited}), including results ChatGPT retrieved but never cited.

Brand visibility — AEO/GEO (25 credits):

GET /v1/chatgpt/brand-visibility?prompt=best+CRM+for+startups&brand=Acme&domain=acme.com&competitors=Salesforce,HubSpot

Required: prompt, brand. Optional: domain, aliases (comma-separated on GET, string[] on POST), competitors (same), country, web_search (default force).

Returns the brand block plus the answer: mentioned, mention_count, first_position (char offset), position_score (1.0 = named at the very start, 0.0 = absent), share_of_voice_pct (brand mentions / brand+competitor mentions), cited, cited_urls[], citation_rank, competitors[] ({name, mentioned, mention_count, first_position, cited, cited_urls}), excerpt, answer, citations[], web_search_triggered, model, country, latency_ms, created_utc/created_at.

List available models (1 credit):

GET /v1/chatgpt/models?country=US

Returns {models: [{slug, title, description, max_tokens, tags[]}], count}. Informational only — /ask does not take a model parameter.

Health (free):

GET /v1/chatgpt/health

Errors come back as {"detail": {"code", "message"}} with codes blocked (422), rate_limited (429), upstream_error, composer_not_found, empty_answer, render_failed, models_unavailable (502). Non-2xx responses are never billed.

Not supported: multi-turn conversations, model selection, streaming, and image input.

---

Perplexity API

Sends a prompt to the real perplexity.ai and returns the grounded answer as structured JSON, including the web sources Perplexity cited. Anonymous — no Perplexity account or API key is involved. Every answer is a live browse taking tens of seconds, so set a generous client timeout and don't poll.

Ask Perplexity (20 credits):

GET /v1/perplexity/ask?prompt=best+open-source+vector+databases&country=US&focus=web

Required: prompt (max 4096 chars). Optional: country (ISO-3166 alpha-2 egress country, default US), focus (web | scholar | social, default web — scholar grounds on academic sources, social on community sources). Also accepts POST with the same fields as a JSON body.

Returns: prompt, answer, answer_markdown, citations[], search_results[], source_domains[], related_questions[], images[]/videos[], web_search_triggered, truncated, search_queries[] (Perplexity's own retrieval fan-out — the sub-queries it issued to its search backend, which no other vendor exposes), focus, model, backend_uuid, country, answer_length, citation_count, latency_ms, created_utc/created_at.

Each Citation is {url, title, snippet, domain, source_type, pub_date_utc, published_at, last_updated, ref_index, start_index, end_index, matched_text} — start_index/end_index are char offsets into answer, so you can anchor a source to the exact sentence it supports. search_results[] is the FULL retrieved set ({url, title, snippet, domain, source_type, pub_date_utc, published_at, last_updated, ref_index, cited}), including results Perplexity retrieved but never cited.

Brand visibility — AEO/GEO (25 credits):

GET /v1/perplexity/brand-visibility?prompt=best+CRM+for+startups&brand=Acme&domain=acme.com&competitors=Salesforce,HubSpot

Required: prompt, brand. Optional: domain, aliases (comma-separated on GET, string[] on POST), competitors (same), country.

Returns the brand block plus the answer: mentioned, mention_count, first_position (char offset), position_score (1.0 = named at the very start, 0.0 = absent), share_of_voice_pct (brand mentions / brand+competitor mentions), cited, cited_urls[], citation_rank, competitors[] ({name, mentioned, mention_count, first_position, cited, cited_urls}), excerpt, answer, citations[], web_search_triggered, model, country, latency_ms, created_utc/created_at.

List available models (1 credit):

GET /v1/perplexity/models?country=US

Returns {models: [{slug, title, description, requires_account}], count}. Informational only — /ask takes a focus parameter, not a model.

Health (free):

GET /v1/perplexity/health

Non-2xx responses are never billed.

Not supported: multi-turn conversations (a backend_uuid is returned but follow-up chaining is not exposed), model selection, streaming, and image input.

---

Baidu API

Web search, news, images and autocomplete from baidu.com — China's #1 search engine (~60% CN market share). No Baidu account or developer key is needed. Chinese and English queries both work; there is no country parameter, Baidu is a single market.

Web search (5 credits):

GET /v1/baidu/search?query=咖啡机&page=1&num=20&language=zh-cn
GET /v1/baidu/search?query=site:zhihu.com+咖啡&time_from=1735689600&time_to=1767225600

Required: query. Optional: page (1-76), num (1-50), language (all|zh-cn|zh-tw), time_from/time_to (unix timestamps). Baidu's own operators — site:, intitle:, inurl:, filetype:, "quoted phrases" — work inside query. page is translated to Baidu's pn offset using num, so the two compose (page=3&num=50 starts at result 101).

Each organic result carries the real target URL in url (from Baidu's mu attribute) alongside baidu_url (the baidu.com/link?url= tracking redirect competitors return on its own), plus display_url, snippet, source, date (Baidu's own string, e.g. 2026年8月1日 / 3天前), date_at (ISO 8601 date), thumbnail, tpl (Baidu's result template id). The response also carries total_results and related_searches[].

News (5 credits):

GET /v1/baidu/news?query=人工智能&sort=time

Required: query. Optional: page (1-76), sort (relevance default, or time for newest first). 10 articles a page; each has title, url, baidu_url, snippet, source, date, date_at, thumbnail.

Images (5 credits):

GET /v1/baidu/images?query=熊猫&page=1

Required: query. Optional: page (1-50). 30 images a page, each with image_url (full size), thumbnail_url, middle_url, hover_url, width, height, type, from_url and from_title (the page it was found on).

Autocomplete (1 credit):

GET /v1/baidu/autocomplete?query=咖啡

Gotchas that will bite you:

• total_results is Baidu's own claim (百度为您找到相关结果约N个) and is not reachable — pagination stops around page 76 regardless.
• Only organic results are returned. Baidu's own vertical blocks (Baike, Tieba, image grids) and ad shells are dropped, so position is organic rank.
• Baidu's SERP carries no time-of-day, so there is no *_utc float — only date (raw) and date_at (ISO date). A midnight timestamp would be fabricated precision.
• url is null when Baidu omits the mu attribute on a result; baidu_url is present either way.
• language and the time window are server-side Baidu filters, not post-filtering — they change which results Baidu returns.
• HTTP 422 is an anti-bot challenge: not billed, and retrying shortly succeeds.

Cost: 1 credit autocomplete; 5 credits search / news / images. health is free.

---

Web Scraping API

Scrape a URL:

POST /v1/web/scrape
Content-Type: application/json

{"url": "https://example.com", "render_js": false, "return_format": "markdown"}

Optional fields: render_js (boolean), wait_for (CSS selector), timeout_ms, proxy_country, return_format (html/markdown/text)

Detect anti-bot protection (1 credit):

POST /v1/web/detect
Content-Type: application/json

{"url": "https://example.com"}

Take screenshot:

POST /v1/web/screenshot
Content-Type: application/json

{"url": "https://example.com", "full_page": true}

Extract structured data:

POST /v1/web/extract
Content-Type: application/json

{"url": "https://example.com", "ai_query": "What is the main product price?"}

Optional fields: extract_rules (CSS/XPath), ai_extract_rules, ai_query

Submit batch job:

POST /v1/web/batch
Content-Type: application/json

{"urls": ["https://example.com", "https://example.org"]}

Check batch status:

GET /v1/web/batch/{job_id}

---

Pagination

Most list endpoints support cursor-based pagination. Pass the cursor query parameter from the previous response to get the next page.

Credits

Each API call costs credits. Check your balance:

• Twitter endpoints: 1 credit per call
• Web scraping: 1-10 credits depending on complexity
• Vinted: 1 credit per call
• Amazon: autocomplete 1, seller profile/feedback 3, listings 5, offers 8, product detail & reviews 10; markets/categories free
• eBay: autocomplete 1; search/completed/item/seller/category 5; reviews/feedback 10; markets/categories free
• Realtor: autocomplete 2; search 8; property detail 10; markets free
• Idealista: suggest 2; search 5; property detail 10; markets free
• Immobiliare: autocomplete 2; search / agency profile / agency listings 5; listing detail & market price insights 10; markets/reference free
• YouTube: autocomplete/oembed 1; resolve/about/subscriber_count/captions 2; search/transcript/channel/playlist/trending/community/shorts/music 5; video detail & batch (per video) 10; streams 15; categories/languages/regions/markets free
• TikTok: oembed 2; profile/video/hashtag/music detail + search + trending + ads 5; list captures (user videos/reposts, comments, replies, related, hashtag/music videos) 8; followers/following/liked + transcript 10; regions free
• ChatGPT: models 1; ask 20; brand-visibility 25; health free
• Perplexity: models 1; ask 20; brand-visibility 25; health free
• Baidu: autocomplete 1; search/news/images 5; health free

Rate Limits

Rate limits are per API key. Check the X-RateLimit-* response headers for current limits.

More Information

• Documentation: https://docs.scrapebadger.com
• Dashboard: https://scrapebadger.com/dashboard?utm_source=openclaw-skill&utm_medium=referral
• MCP Server: https://mcp.scrapebadger.com/mcp