# ScrapeCrate > ScrapeCrate is one REST API for public web data: YouTube API, Google Maps API, Trustpilot Reviews API, Meta Ad Library API. One account, one API key and one credit balance for every API. 100 free credits every month, then $1 per 1,000 credits; a call costs YouTube 1, Google Maps 3, Meta Ad Library 5, Trustpilot 8 credits, purchased credits are valid for 12 months and only successful calls are billed. Short index: https://scrapecrate.com/llms.txt. Human docs: https://scrapecrate.com/docs. OpenAPI 3.1: https://api.scrapecrate.com/v1/openapi.json. Catalog JSON: https://api.scrapecrate.com/v1/catalog. ## Quickstart 1. Sign up at https://scrapecrate.com/auth/sign-up (no credit card) and create an API key in the dashboard (https://scrapecrate.com/dashboard/api-keys). The key is shown once. 2. Call any endpoint under the base URL `https://api.scrapecrate.com/v1` with the key in the `X-API-Key` header or as `Authorization: Bearer `. Every endpoint is GET and answers JSON. 3. Read `data` (object for lookups, array for lists), `pagination` (lists only) and `meta` (cost and balance). ```bash curl "https://api.scrapecrate.com/v1/youtube/transcripts/dQw4w9WgXcQ" -H "X-API-Key: YOUR_API_KEY" ``` ## Conventions (every API) - Base URL: `https://api.scrapecrate.com/v1` (not an endpoint itself). The endpoint paths below are written from the API host, `/v1` included. - Authentication: `X-API-Key: ` or `Authorization: Bearer `. Never in the query string. - Success body: `{"data": ..., "pagination": {"next_cursor": string|null, "has_more": bool} (lists only), "meta": {"request_id", "credits_used", "credits_remaining"}}`. - Pagination: a call returns at most one page (the `limit` maximum is one page; above it: 400 `invalid_parameter`). Pass `pagination.next_cursor` as the `cursor` parameter with the same other parameters; every page is a new call; stop when `has_more` is false. Expired cursor: 410 `cursor_expired`. - Headers: `X-Request-Id` (every response), `X-Credits-Used` and `X-Credits-Remaining` (successful calls, and failed calls after the balance check, with 0 used), `Retry-After` (429 and some 503). - Rate limits: 300 requests per minute and 20 concurrent requests per API key. - `GET /v1/account`: the balance of the key's account (`credits_remaining`, `free_credits_remaining`, `purchased_credits_remaining`, `free_monthly_credits`). Free of charge. ## Errors Every error (status >= 400) has this body; `details` is present only on some errors. Branch on `error.code`, not on the message. `request_id` is `req_none` when the request was rejected before reaching an endpoint (unknown path, method other than GET, malformed request). Failed calls are never billed. ```json { "error": { "code": "invalid_parameter", "message": "Some parameters are not valid.", "request_id": "req_8f2c1a9b0d3e4f5a6b7c", "details": [{ "param": "limit", "message": "must be <= 100" }] } } ``` | Status | Code | Meaning | Retry | |---|---|---|---| | 400 | `invalid_parameter` | A parameter is missing, unknown or invalid. details lists each problem as {"param", "message"}, e.g. "is required", "must be <= 100", "unknown parameter". | No: fix the request | | 400 | `invalid_cursor` | The cursor is not valid for this endpoint. | No | | 401 | `missing_api_key` | No API key in the X-API-Key header or the Authorization: Bearer header. | No | | 401 | `invalid_api_key` | The key is wrong, disabled or deleted. | No | | 402 | `insufficient_credits` | Your balance is below the minimum price of the call. details is {"required", "available"}. | After a top-up | | 403 | `account_disabled` | The account is suspended. | No: contact support | | 404 | `not_found, *_not_found` | The video, place, company, ad... does not exist or is not public (an API can use a more specific code ending in _not_found). Also an unknown endpoint path. | No | | 405 | `method_not_allowed` | Every endpoint is GET. | No | | 410 | `cursor_expired` | The cursor is too old: start again from the first page. | Restart the list | | 429 | `rate_limited` | Over 300 requests per minute for the key (or over the per-IP limit). | Yes, after Retry-After | | 429 | `too_many_concurrent_requests` | More than 20 requests in flight for the key. | Yes, after Retry-After | | 429 | `insufficient_credits_retry_later` | The key kept retrying with no credits left, so its requests are paused for a while. details is {"retry_after"}; a top-up lifts the pause within seconds. | After a top-up, or after Retry-After | | 500 | `internal_error` | Unexpected error on our side. | Yes, with backoff | | 501 | `not_available` | This endpoint is not available yet. | No | | 502 | `source_error` | The data source returned an unexpected answer. | Yes, with backoff | | 503 | `not_available` | This API is not available yet (announced, not launched). Nothing is charged. | No: it will not work until launch | | 503 | `source_unavailable` | The data source is temporarily unavailable. | Yes, with backoff (Retry-After when present) | | 503 | `source_busy` | The data source is saturated. | Yes, after Retry-After | | 504 | `source_timeout` | The data source took too long to answer. | Yes | Full page: https://scrapecrate.com/docs/errors ## Pricing - 1 credit = $0.001 at the pay-as-you-go rate ($1 per 1,000 credits). 100 free credits every month (they reset monthly and are used first). - Credits per successful call, the same on every endpoint of an API and whatever it returns: YouTube 1, Google Maps 3, Meta Ad Library 5, Trustpilot 8. YouTube comments: 1 credit per 100 comments. An empty list is a successful call. Only successful (200) calls cost credits. - In calls: 100 free credits = 100 YouTube calls, or 33 Google Maps, 20 Meta Ad Library, 12 Trustpilot calls a month; 22,000 credits = 22k YouTube calls, or 7.3k Google Maps, 4.4k Meta Ad Library, 2.75k Trustpilot calls. - Purchased and subscription credits are valid for 12 months from purchase (credits bought before this rule never expire); the ones that expire first are used first, and an email warns 30 days before. They work on every API. The value of a credit never changes; credits per call may change, announced on the pricing page. - Top-up: from $5 to $10,000. - Credit packs: 10K credits for $9; 100K credits for $80; 1M credits for $700; 10M credits for $5,000. - Monthly plans: Starter 22,000 credits for $19/month; Growth 60,000 credits for $49/month; Scale 200,000 credits for $149/month. - Price list: https://scrapecrate.com/pricing ## SDKs - Python (3.9+): `pip install scrapecrate`. `from scrapecrate import ScrapeCrate; client = ScrapeCrate(api_key="...")` (or set `SCRAPECRATE_API_KEY`). Async client `AsyncScrapeCrate`. `client..(**params)`, `client..iterate("", **params)` follows cursors. - TypeScript / JavaScript (Node 18+, Deno, Bun): `npm install scrapecrate`. `new ScrapeCrate({ apiKey })`, `await client..({...})`, `for await (const x of client..iterate('', {...}))`. - Both retry 429/500/502/503/504 up to 3 times with backoff and raise typed errors: ValidationError (400), AuthenticationError (401), InsufficientCreditsError (402), PermissionDeniedError (403), NotFoundError (404), RateLimitError (429), ServiceUnavailableError (5xx), all subclasses of ApiError with status, code, message, request id and details. - Docs: https://scrapecrate.com/docs/sdks ## MCP server - Run with `uvx scrapecrate-mcp` (Python 3.10+), with the key in `SCRAPECRATE_API_KEY`. Claude Code: `claude mcp add scrapecrate -e SCRAPECRATE_API_KEY=YOUR_API_KEY -- uvx scrapecrate-mcp`. - Client config: `{"mcpServers": {"scrapecrate": {"command": "uvx", "args": ["scrapecrate-mcp"], "env": {"SCRAPECRATE_API_KEY": "YOUR_API_KEY"}}}}` - One tool per endpoint named `_` (e.g. `youtube_transcript`, `maps_place_reviews`), plus `account_balance` (free). Billed like API calls; failed calls are free. - Docs: https://scrapecrate.com/docs/mcp ## YouTube API Fetch transcripts in every available language, video metadata and chapters, comments, channel details and video lists, and keyword search, as clean JSON. Overview: https://scrapecrate.com/youtube-api. Reference: https://scrapecrate.com/docs/api/youtube. ### GET /v1/youtube/transcripts/{video_id} Get a transcript. The transcript of a video as timed segments and plain text. Without `lang` you get the video's default caption track. - Price: 1 credit / call ($1 per 1,000 calls). - SDK: Python `client.youtube.transcript()`, TypeScript `client.youtube.transcript()`; MCP tool `youtube_transcript`. - Parameters: - `video_id` (path, required, string): 11-character video ID - `lang` (query, optional, string): Language code (e.g. en, it, pt-BR); auto-translated when the video doesn't have it - `format` (query, optional, one of json, text, srt, default json): json (segments + text), text or srt - Example: `https://api.scrapecrate.com/v1/youtube/transcripts/dQw4w9WgXcQ` ### GET /v1/youtube/transcripts/{video_id}/languages List transcript languages. The caption tracks available for a video, manual and auto-generated. - Price: 1 credit / call ($1 per 1,000 calls). - SDK: Python `client.youtube.transcript_languages()`, TypeScript `client.youtube.transcriptLanguages()`; MCP tool `youtube_transcript_languages`. - Parameters: - `video_id` (path, required, string): 11-character video ID - Example: `https://api.scrapecrate.com/v1/youtube/transcripts/dQw4w9WgXcQ/languages` ### GET /v1/youtube/videos/{video_id} Get video metadata. Title, description, channel, duration, views, likes, comment count, tags and categories. - Price: 1 credit / call ($1 per 1,000 calls). - SDK: Python `client.youtube.video()`, TypeScript `client.youtube.video()`; MCP tool `youtube_video`. - Parameters: - `video_id` (path, required, string): 11-character video ID - Example: `https://api.scrapecrate.com/v1/youtube/videos/dQw4w9WgXcQ` ### GET /v1/youtube/videos/{video_id}/chapters Get video chapters. The chapters of a video with start and end times in seconds. - Price: 1 credit / call ($1 per 1,000 calls). - SDK: Python `client.youtube.video_chapters()`, TypeScript `client.youtube.videoChapters()`; MCP tool `youtube_video_chapters`. - Parameters: - `video_id` (path, required, string): 11-character video ID - Example: `https://api.scrapecrate.com/v1/youtube/videos/dQw4w9WgXcQ/chapters` ### GET /v1/youtube/videos/{video_id}/comments List video comments. Top-level comments of a video with author, likes, replies count, date and pinned flag. - Price: 1 credit / 100 items ($0.010 per 1,000 items). Paginated with `cursor`: each page is a new call. - SDK: Python `client.youtube.video_comments()`, TypeScript `client.youtube.videoComments()`; MCP tool `youtube_video_comments`. - Parameters: - `video_id` (path, required, string): 11-character video ID - `sort` (query, optional, one of top, new, default top): Top comments or newest first - `limit` (query, optional, integer, 1-100, default 50): Comments per page (max 100) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/youtube/videos/dQw4w9WgXcQ/comments` ### GET /v1/youtube/channels/{channel_id} Get a channel. Channel details by channel ID (`UC...`) or @handle. - Price: 1 credit / call ($1 per 1,000 calls). - SDK: Python `client.youtube.channel()`, TypeScript `client.youtube.channel()`; MCP tool `youtube_channel`. - Parameters: - `channel_id` (path, required, string): Channel ID (UC...) or @handle - Example: `https://api.scrapecrate.com/v1/youtube/channels/%40RickAstleyYT` ### GET /v1/youtube/channels/{channel_id}/videos List channel videos. The videos of a channel, newest first, page by page. - Price: 1 credit / call ($1 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.youtube.channel_videos()`, TypeScript `client.youtube.channelVideos()`; MCP tool `youtube_channel_videos`. - Parameters: - `channel_id` (path, required, string): Channel ID (UC...) or @handle - `include_shorts` (query, optional, boolean, default false): Include Shorts - `limit` (query, optional, integer, 1-100, default 30): Videos per page (max 100) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/youtube/channels/%40RickAstleyYT/videos` ### GET /v1/youtube/search Search videos. Search by keyword and get videos with channel, duration, views and upload date. - Price: 1 credit / call ($1 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.youtube.search()`, TypeScript `client.youtube.search()`; MCP tool `youtube_search`. - Parameters: - `q` (query, required, string): Search query - `limit` (query, optional, integer, 1-50, default 20): Results (max 50) - Example: `https://api.scrapecrate.com/v1/youtube/search?q=lofi%20hip%20hop` ## Google Maps API Search businesses and places by keyword and area, get address, phone, website, hours, rating and attributes of a place, and page through its reviews with owner responses. Overview: https://scrapecrate.com/google-maps-api. Reference: https://scrapecrate.com/docs/api/maps. ### GET /v1/maps/places/search Search places. Places matching a query, optionally around a point, up to 20 per call. - Price: 3 credits / call ($3 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.maps.places_search()`, TypeScript `client.maps.placesSearch()`; MCP tool `maps_places_search`. - Parameters: - `q` (query, required, string): What to search, e.g. dentist in Berlin - `language` (query, optional, string, default en): Result language (en, it, pt-BR...) - `region` (query, optional, string, default us): Two-letter country code to bias results - `lat` (query, optional, number, -90-90): Latitude of the search centre (with lng) - `lng` (query, optional, number, -180-180): Longitude of the search centre (with lat) - `radius_km` (query, optional, number, 0.1-100): Radius around lat/lng in km - `limit` (query, optional, integer, 1-20, default 20): Places per page (max 20) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/maps/places/search?q=pizzeria%20in%20Milan` ### GET /v1/maps/places/{place_id} Get place details. Full details of a place by place ID, CID or feature ID. - Price: 3 credits / call ($3 per 1,000 calls). - SDK: Python `client.maps.place()`, TypeScript `client.maps.place()`; MCP tool `maps_place`. - Parameters: - `place_id` (path, required, string): Place ID, CID or feature ID (use /places/resolve for URLs) - `language` (query, optional, string, default en): Result language - `region` (query, optional, string, default us): Two-letter country code - Example: `https://api.scrapecrate.com/v1/maps/places/ChIJ2eUgeAK6j4ARbn5u_wAGqWA` ### GET /v1/maps/places/{place_id}/reviews List place reviews. Reviews of a place with rating, text, translation, date and owner response. - Price: 3 credits / call ($3 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.maps.place_reviews()`, TypeScript `client.maps.placeReviews()`; MCP tool `maps_place_reviews`. - Parameters: - `place_id` (path, required, string): Place ID, CID or feature ID - `sort` (query, optional, one of relevant, newest, highest, lowest, default relevant): Review order - `language` (query, optional, string, default en): Language of translations - `limit` (query, optional, integer, 1-20, default 20): Reviews per page (max 20, one page of the source) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/maps/places/ChIJ2eUgeAK6j4ARbn5u_wAGqWA/reviews` ### GET /v1/maps/places/resolve Resolve a Maps URL or ID. Turn any Google Maps URL, short link, CID or feature ID into the place IDs. - Price: 3 credits / call ($3 per 1,000 calls). - SDK: Python `client.maps.places_resolve()`, TypeScript `client.maps.placesResolve()`; MCP tool `maps_places_resolve`. - Parameters: - `input` (query, required, string): A Maps URL, short link, CID or feature ID - Example: `https://api.scrapecrate.com/v1/maps/places/resolve?input=https%3A%2F%2Fmaps.app.goo.gl%2Fabc123` ## Trustpilot Reviews API Search companies, get their trust score, star distribution and reply behaviour, and download their reviews with filters on stars, language, dates and verification. Overview: https://scrapecrate.com/trustpilot-api. Reference: https://scrapecrate.com/docs/api/trustpilot. ### GET /v1/trustpilot/companies/search Search companies. Companies matching a name or domain. - Price: 8 credits / call ($8 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.trustpilot.companies_search()`, TypeScript `client.trustpilot.companiesSearch()`; MCP tool `trustpilot_companies_search`. - Parameters: - `query` (query, required, string): Company name or domain - `limit` (query, optional, integer, 1-10, default 10): Companies per page (max 10, one page of the source) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/trustpilot/companies/search?query=booking.com` ### GET /v1/trustpilot/companies/{domain} Get a company. Company profile with trust score, star distribution, languages and reply behaviour. - Price: 8 credits / call ($8 per 1,000 calls). - SDK: Python `client.trustpilot.company()`, TypeScript `client.trustpilot.company()`; MCP tool `trustpilot_company`. - Parameters: - `domain` (path, required, string): Company domain, e.g. booking.com - `include_similar` (query, optional, boolean, default false): Include similar companies - Example: `https://api.scrapecrate.com/v1/trustpilot/companies/booking.com` ### GET /v1/trustpilot/companies/{domain}/reviews List company reviews. Public reviews of a company with filters. `meta.complete` is false when the source caps the reviews reachable for a filter combination. - Price: 8 credits / call ($8 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.trustpilot.company_reviews()`, TypeScript `client.trustpilot.companyReviews()`; MCP tool `trustpilot_company_reviews`. - Parameters: - `domain` (path, required, string): Company domain - `stars` (query, optional, string): Comma-separated star ratings - `language` (query, optional, string, default all): all or an ISO 639-1 code - `since` (query, optional, date): Published on or after (YYYY-MM-DD) - `until` (query, optional, date): Published on or before (YYYY-MM-DD) - `verified` (query, optional, boolean, default false): Only verified reviews - `with_reply` (query, optional, boolean, default false): Only reviews with a company reply - `search` (query, optional, string): Full-text filter - `sort` (query, optional, one of recent, relevant, default recent): Order - `limit` (query, optional, integer, 1-20, default 20): Reviews per page (max 20, one page of the source) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/trustpilot/companies/booking.com/reviews` ### GET /v1/trustpilot/categories List categories. The category tree used to browse companies. - Price: 8 credits / call ($8 per 1,000 calls). - SDK: Python `client.trustpilot.categories()`, TypeScript `client.trustpilot.categories()`; MCP tool `trustpilot_categories`. - Example: `https://api.scrapecrate.com/v1/trustpilot/categories` ### GET /v1/trustpilot/categories/{category_id}/companies List companies in a category. Companies of a category in a country, with optional trust score and city filters. - Price: 8 credits / call ($8 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.trustpilot.category_companies()`, TypeScript `client.trustpilot.categoryCompanies()`; MCP tool `trustpilot_category_companies`. - Parameters: - `category_id` (path, required, string): Category ID from /categories - `country` (query, required, string): Two-letter country code - `sort` (query, optional, one of relevant, reviews_count, default relevant): Order - `min_trust_score` (query, optional, one of 3.0, 4.0, 4.5): Minimum trust score - `location` (query, optional, string): City, e.g. Milano - `limit` (query, optional, integer, 1-20, default 20): Companies per page (max 20, one page of the source) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/trustpilot/categories/travel_agency/companies?country=IT` ## Meta Ad Library API Search active and past ads by keyword or advertiser across countries, get creatives, copy, platforms and run dates, and EU transparency data of a single ad. Overview: https://scrapecrate.com/meta-ad-library-api. Reference: https://scrapecrate.com/docs/api/ads. ### GET /v1/ads/search Search ads. Ads matching a keyword or an advertiser page, with filters, up to 30 per call. - Price: 5 credits / call ($5 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.ads.ads_search()`, TypeScript `client.ads.adsSearch()`; MCP tool `ads_ads_search`. - One of `q`, `page_id` is required. - Parameters: - `q` (query, optional, string): Keyword (required unless page_id is set) - `page_id` (query, optional, string): Only ads of this advertiser page - `country` (query, optional, string, default ALL): Two-letter country code or ALL - `status` (query, optional, one of active, inactive, all, default active): Ad status - `ad_type` (query, optional, one of all, political, default all): All ads or political/issue ads - `media_type` (query, optional, one of all, image, video, meme, none, default all): Creative type - `platforms` (query, optional, string): Comma-separated: facebook, instagram, messenger, audience_network - `languages` (query, optional, string): Comma-separated ISO 639-1 codes - `active_from` (query, optional, date): Running on or after (YYYY-MM-DD) - `active_to` (query, optional, date): Running on or before (YYYY-MM-DD) - `match` (query, optional, one of any, exact, default any): Keyword match - `sort` (query, optional, one of relevance, impressions, default relevance): Order - `limit` (query, optional, integer, 1-30, default 30): Ads per page (max 30, one page of the source) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/ads/search?q=running%20shoes` ### GET /v1/ads/{ad_id} Get an ad. One ad with creatives, payer and beneficiary, and EU/UK transparency data. - Price: 5 credits / call ($5 per 1,000 calls). - SDK: Python `client.ads.ad()`, TypeScript `client.ads.ad()`; MCP tool `ads_ad`. - Parameters: - `ad_id` (path, required, string): Ad Library ID - `country` (query, optional, string, default ALL): Country used to look the ad up - Example: `https://api.scrapecrate.com/v1/ads/1234567890123456` ### GET /v1/ads/pages/search Search advertisers. Advertiser pages whose ads match a keyword. - Price: 5 credits / call ($5 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.ads.pages_search()`, TypeScript `client.ads.pagesSearch()`; MCP tool `ads_pages_search`. - Parameters: - `q` (query, required, string): Keyword - `country` (query, optional, string, default ALL): Two-letter country code or ALL - `limit` (query, optional, integer, 1-10, default 10): Pages (max 10, one page of the source) - Example: `https://api.scrapecrate.com/v1/ads/pages/search?q=running%20shoes` ### GET /v1/ads/pages/{page_id} Get an advertiser. Advertiser page details with Instagram account and active ads estimate. - Price: 5 credits / call ($5 per 1,000 calls). - SDK: Python `client.ads.page()`, TypeScript `client.ads.page()`; MCP tool `ads_page`. - Parameters: - `page_id` (path, required, string): Advertiser page ID - Example: `https://api.scrapecrate.com/v1/ads/pages/102030405060708` ### GET /v1/ads/pages/{page_id}/ads List an advertiser's ads. The ads of one advertiser page, with the same filters as search. - Price: 5 credits / call ($5 per 1,000 calls). Paginated with `cursor`: each page is a new call. - SDK: Python `client.ads.page_ads()`, TypeScript `client.ads.pageAds()`; MCP tool `ads_page_ads`. - Parameters: - `page_id` (path, required, string): Advertiser page ID - `country` (query, optional, string, default ALL): Two-letter country code or ALL - `status` (query, optional, one of active, inactive, all, default active): Ad status - `media_type` (query, optional, one of all, image, video, meme, none, default all): Creative type - `limit` (query, optional, integer, 1-30, default 30): Ads per page (max 30, one page of the source) - `cursor` (query, optional, string): pagination.next_cursor of the previous page - Example: `https://api.scrapecrate.com/v1/ads/pages/102030405060708/ads`