API reference

Up to 100 full public TikTok video URLs. Use Profile usernames below for profiles; invalid video URLs are skipped independently. Default: [].

Maximum total delivered comments and replies for each video. Default: 100. Range: 1–10000.

Maximum top-level comments collected for each video. Default: 100. Range: 1–10000.

Per-parent reply upper bound, default 0. Delivers main comments first; explicitly set this to collect replies. Default: 0. Range: 0–500.

Up to 100 usernames, optional @. Videos and public reposts supported. Default: [].

Default 1. Upper bound on discovered posts per profile, not comment rows. Default: 1. Range: 1–100.

Default latest. Native video discovery order: latest, popular or oldest. Reposts retain upstream repost order. Default: "latest".

Default [videos]. Select videos, reposts or both. Each section uses native upstream pagination; profileSorting applies to videos. resultsPerPage caps unique posts across selected sections. Default: ["videos"].

Inclusive UTC publication lower bound: valid YYYY-MM-DD or integer string 1..36500 (1=today, 2=yesterday). Relative days freeze at each profile unit first request, before discovery.

YYYY-MM-DD UTC video publication date; inclusive calendar day. Applies to profile discovery before comment collection.

Default false. Profile discovery only. Default: false.

Process 1–16 video or profile targets at once. Each profile processes its videos sequentially. Output can arrive in completion order; this is not a requests-per-second setting. Default: 16. Range: 1–16.

Verified comment ID.

Public comment or reply text; null if unavailable.

Unix seconds when supplied.

Corresponding ISO UTC creation time.

Source-reported comment likes; null is unknown.

Source-reported reply count, not a completeness guarantee.

Public author ID when available.

Public author username when available.

Null for top-level comments; directed reply target when supplied, otherwise parent. Target may be outside the sample.

Public author image URL reference; may expire.

Verified source flag when supplied.

Verified source pin flag, or null when unknown.

Empty array for verified absence; otherwise null when handles cannot be verified.

Verified Tag-contract identities with id; name, nickName and profileUrl remain null. Unknown or unverified metadata stays null.

Canonical full TikTok video URL.

Actual video URL used for collection, including profile-discovered videos.

Actual video input URL.

Protected contractVersion, inputHash and video-ID:comment-ID recovery identity; no cookies or credentials.

Up to 100 public profile URLs or handles. First occurrences are kept. Invalid entries are skipped. Takes precedence over handle, including when empty.

Compatibility input used only when profileUrls is absent. Accepts a bare handle or @handle.

Positive integer profile ID supplied by the public OnlyFans source.

Public profile username.

Public profile display name.

Canonical OnlyFans profile URL.

Public biography; may contain source HTML.

Verification flag from the public source.

Optional anonymous source flag; preserves true and false.

Public profile location if supplied.

Public website if supplied; not independently verified.

Source-reported public post count.

Source-reported public photo count.

Source-reported public video count.

Source-reported public media count; no media is downloaded.

Public subscriber count when supplied; null does not mean zero.

Source public subscription price; no currency is inferred.

Optional validated source subscription offers.

Optional validated source promotions.

Profile image URL reference; may expire.

Profile header image URL reference; may expire.

Optional available source-provided social URL; not independently verified and never inferred.

Optional available source-provided social URL; not independently verified and never inferred.

Optional available source-provided social URL; not independently verified and never inferred.

Optional available source-provided social URL; not independently verified and never inferred.

Public YouTube channel handles, IDs or URLs. Empty input succeeds with no results or result charges.

Legacy mode: processed in this order under a shared maxResults cap. Independent limits override tabs when maxResultsShorts or maxResultStreams is supplied. Default: ["videos"].

Uses YouTube native category ordering; unsupported orders are reported as category errors.

Legacy mode: shared maximum per channel (1–1000). Independent mode: videos maximum (0–1000), enabled by maxResultsShorts or maxResultStreams. Duplicates across channels are delivered once. Default: 30. Range: 0–1000.

Additional public watch request only for selected items. Failed enrichment retains the basic row and descriptionComplete=false. Default: false.

Optional: supplying this activates independent category limits. Unspecified Shorts and Streams limits are zero. Set all three limits to zero for channel-only output. Range: 0–1000.

Optional: supplying this activates independent category limits. Unspecified Shorts and Streams limits are zero. Set all three limits to zero for channel-only output. Range: 0–1000.

Maximum simultaneous selected-item detail requests. Results retain channel-list order. Default: 8. Range: 1–32.

Optional YYYY-MM-DD (UTC day boundary) or N day / N days (rolling run-start cutoff). Forces latest order. Unknown and boundary-uncertain dates are retained and reported in OUTPUT.dateFilter.

Fetch channel About once per channel. Profile fields appear in OUTPUT.channels and CHANNELS.json; unavailable fields remain null. Adds up to one upstream request per channel. Default: false.

Alternative to channels. Public YouTube channel URLs only; cannot be combined with channels.

Alternative to sort; cannot be combined with sort.

streamers writes COMPATIBILITY.json using competitor field names plus explicit precision and semantics. Dataset remains v1. Default: "v1".

Source identity; YouTube IDs identify videos.

Canonical YouTube video or Shorts URL.

Public video title.

video, short or stream.

Available public channel ID.

Available public channel URL.

Available public channel title.

Full public description if requested and available; otherwise null or available listing evidence.

Available source calendar date; not an inferred exact timestamp.

Always null; exact timestamps are not inferred from relative labels.

Original relative or public display date label.

Source evidence for displayed date, or unknown.

Original displayed view count, including abbreviations.

Count precision evidence; abbreviated_or_unparsed values must not be treated as exact counts.

Original displayed duration.

Public image URL reference; may expire.

Submitted channel reference associated with this row.

Exact count only when supported by source evidence; abbreviations remain null here.

Available duration in seconds.

True only for a complete verified description.

Native Dataset contract version: 1.

ok, unavailable or not_requested; failed optional detail preserves the listing row.

Requested ordering.

confirmed, requested_unconfirmed or null. Unsupported sorts are reported as category errors.

live, upcoming or null; concurrent audience differs from accumulated views.

Job title, skills, or free-text search (e.g. "software engineer remote"). Recommended for new searches; ignored when search URLs are provided.

City, region, or country as you would type it on LinkedIn (e.g. "United States", "London, United Kingdom", "Remote").

Only return jobs posted within this window. Default: "anyTime". Accepted: "anyTime", "past24Hours", "pastWeek", "pastMonth".

Maximum number of jobs to scrape from each input URL (or from the structured filters when no URLs are provided). Defaults to 1000 when omitted. Default: 1000. Minimum: 1.

Add the company website, description, LinkedIn employee count, address, slogan, and logo. Enabled by default and automatically cached for faster repeated runs. Default: true.

Optional LinkedIn geoId for precise location targeting (found in a LinkedIn search URL as `geoId=`). Overrides free-text location when LinkedIn resolves it.

Search radius in miles around the location. Leave empty for LinkedIn's default. Minimum: 0.

LinkedIn numeric company IDs (from a company URL or `f_C=` in a search URL), e.g. "1441" for Google. Default: [].

Only jobs with fewer than 10 applicants. Default: false.

Advanced: paste one or more public LinkedIn Jobs search URLs to preserve an existing workflow. When URLs are provided, their search criteria take priority over the structured filters. Leave this empty for the recommended keyword search. Default: [].

Enable this to split your search by cities within a country. This helps bypass LinkedIn's 1000 job limit per search URL by creating separate searches for each city. This will overwrite the location filter in input search URLs. Default: false.

Select the country whose cities will be used to split the search. Required only when 'Split search by city locations' is enabled. Accepted: "US", "CA", "MX", "GB", "DE", "FR", "NL", "CH", "SE", "NO", "DK", "FI", "IE", "ES", "IT", "PT", "BE", "AT", "PL", "CZ", "RO", "HU", "GR", "BG", "HR", "RS", "UA", "SK", "LT", "LV", "EE", "SI", "LU", "MT", "IS", "CY", "TR", "RU", "GE", "AM", "AZ", "BY", "BA", "XK", "MD", "ME", "MK", "AL", "MC", "AD", "LI", "SM", "VA", "IN", "CN", "JP", "KR", "SG", "HK", "TW", "TH", "VN", "ID", "MY", "PH", "BD", "PK", "LK", "NP", "MM", "KH", "KZ", "UZ", "KG", "TJ", "TM", "MN", "BN", "MV", "BT", "LA", "TL", "AF", "KP", "AE", "SA", "IL", "QA", "KW", "BH", "OM", "JO", "LB", "IQ", "IR", "PS", "SY", "YE", "AU", "NZ", "FJ", "PG", "WS", "TO", "VU", "SB", "KI", "MH", "FM", "PW", "NR", "TV", "BR", "AR", "CO", "CL", "PE", "EC", "UY", "VE", "BO", "PY", "GY", "SR", "CR", "PA", "GT", "DO", "PR", "JM", "TT", "HN", "SV", "NI", "BS", "BB", "CU", "HT", "BZ", "AG", "DM", "GD", "LC", "KN", "VC", "ZA", "NG", "KE", "EG", "MA", "GH", "ET", "TZ", "RW", "UG", "SN", "TN", "CI", "CM", "DZ", "AO", "MU", "ZM", "ZW", "MZ", "CD", "BW", "NA", "MG", "LY", "SD", "SS", "SO", "ML", "GA", "BJ", "BF", "MW", "NE", "TD", "SL", "LR", "ER", "GM", "GN", "GW", "TG", "MR", "SZ", "LS", "DJ", "CG", "CF", "BI", "CV", "KM", "GQ", "ST", "SC".

Preserve filters from older LinkedIn search URLs with LinkedIn's current AI search. Experience level, job type, and workplace filters are converted into search terms. Supported URL filters remain unchanged. Other ID-only filters may not be interpreted by LinkedIn; describe them in Keywords instead. Default: true.

LinkedIn job identifier.

Public LinkedIn job URL.

Job title.

Hiring company name.

Public LinkedIn company URL when available.

Public company logo URL when available.

Location shown on the job posting.

Published salary text when available.

Published posting date when available.

Published benefits when available.

Full job description with HTML formatting when available.

Applicant count text when publicly available.

Public application URL when available.

Full plain-text job description when available.

Publicly listed job poster name.

Publicly listed job poster title.

Public job poster photo URL.

Public job poster profile URL.

Seniority level shown on the job posting when available.

Employment type shown on the job posting when available.

Job function shown on the job posting when available.

Industries shown on the job posting when available.

Public company description.

Public company website URL.

LinkedIn employee count when available.

Structured public company address when available.

Public company slogan when available.

Search source that produced the job.

LinkedIn search reference identifier when available.

LinkedIn search tracking identifier when available.

LinkedIn company URLs or slugs. Valid company and Showcase pages are supported. Default: [].

Optional best-effort exact-name resolution. For reliable bulk delivery, use LinkedIn company URLs or slugs. Default: [].

Maximum number of unique input companies to process. Invalid or unresolved inputs may reduce the number of delivered records. Default: 100. Minimum: 1. Maximum: 1000.

Maximum number of company batches and name resolutions processed in parallel. Default: 3. Minimum: 1. Maximum: 10.

Company name

Canonical LinkedIn company URL

Company website, when publicly exposed.

Public company description.

Primary public industry label.

Public company type.

Public employee range label.

Employees shown on LinkedIn, when available.

Public LinkedIn follower count, when available.

Year the company was founded, when available.

Public headquarters display text.

Street address, when available.

City, when available.

State or region, when available.

Country, when available.

Postal code, when available.

Public company specialties.

Public company logo URL, when available.

Original company name input, when name resolution was used.

Name resolution confidence; currently high for exact matches.

Source used to resolve a company name.

UTC scrape timestamp.

One to 1,000 public LinkedIn post URLs or numeric activity IDs. Duplicate identities are fetched once and reported explicitly at their later input positions.

True only for a successfully fetched public post.

Original input value.

LinkedIn activity ID.

Canonical public post URL.

Post text exposed publicly.

Observed public post type.

Exact JSON-LD publication time when exposed.

Public author information.

Public images, video, or document resources.

Public external article card.

Public attached job card.

Public original-post details fetched from its own page.

Only public total reactions and comments.

Live public-source declaration.

Time of the live fetch.

Unavailable public-data capabilities.

Explicit non-billable invalid, restricted, or duplicate result.

One record per known person. A full name is required; company, domain, title, location, and keywords improve ranking.

Maximum ranked LinkedIn profile candidates returned for each successful input. Default: 5. Minimum: 1. Maximum: 10.

Include Google result titles, snippets, and match reasons with each candidate. Default: true.

Higher thresholds return fewer candidates. A score is ranking evidence, not an identity-verification guarantee. Default: 40. Minimum: 0. Maximum: 100.

Maximum people processed concurrently. The conservative default is validated for the Google-only route. Default: 5. Minimum: 1. Maximum: 10.

Input person's full name.

Optional input company.

Optional input company domain.

Optional input role or title.

Optional input location.

Optional input keywords.

Candidate, no reliable indexed result, or execution/input error. Accepted: "candidate_found", "not_found", "error".

Top canonical linkedin.com/in candidate, or null.

Google-indexed title for the top candidate.

Google-indexed snippet for the top candidate.

Deterministic ranking score from public input and Google evidence.

Ranking band, never an identity guarantee. Accepted: "high", "medium", "low", "none".

Public clues that contributed to the ranking score.

Always true because candidate identity must be reviewed.

Public Google-indexed web results. Accepted: "google_public_index".

Google-only query variants actually executed.

Ranked, deduplicated canonical LinkedIn profile candidates.

Machine-readable error code. Accepted: "invalid_input", "search_failed".

User-readable error details.

What to look for, one term per line. "dentist in Austin TX" style terms with a place name work best. Each term is searched separately and results keep the order you typed.

A single place to search in, for example "Austin, TX". Combined with each search term. Leave empty if your terms already name the place.

Up to 20. Google returns one page of 20 per search, and this version does not page further, so higher values are capped at 20. Use more search terms or a tighter location to get more businesses. Default: 20. Minimum: 1. Maximum: 20.

Language of names, categories and opening hours. Default: "en". Accepted: "en", "af", "az", "id", "ms", "bs", "ca", "cs", "da", "de", "et", "es", "es-419", "eu", "fil", "fr", "gl", "hr", "zu", "is", "it", "sw", "lv", "lt", "hu", "nl", "no", "uz", "pl", "pt-BR", "pt-PT", "ro", "sq", "sk", "sl", "fi", "sv", "vi", "tr", "el", "bg", "ky", "kk", "mk", "mn", "ru", "sr", "uk", "ka", "hy", "iw", "ur", "ar", "fa", "am", "ne", "hi", "mr", "bn", "pa", "gu", "ta", "te", "kn", "ml", "si", "th", "lo", "my", "km", "ko", "ja", "zh-CN", "zh-TW".

Only returns (and charges for) businesses whose address country or time zone confirms they are in this country. Businesses whose country cannot be confirmed are left out. Country detection relies on Google's data and is not guaranteed to be perfect. Leave empty to search without a country filter. Accepted: "", "us", "af", "al", "dz", "as", "ad", "ao", "ai", "aq", "ag", "ar", "am", "aw", "au", "at", "az", "bs", "bh", "bd", "bb", "by", "be", "bz", "bj", "bm", "bt", "bo", "ba", "bw", "bv", "br", "io", "bn", "bg", "bf", "bi", "kh", "cm", "ca", "cv", "ky", "cf", "td", "cl", "cn", "cx", "cc", "co", "km", "cg", "cd", "ck", "cr", "ci", "hr", "cu", "cy", "cz", "dk", "dj", "dm", "do", "ec", "eg", "sv", "gq", "er", "ee", "et", "fk", "fo", "fj", "fi", "fr", "gf", "pf", "tf", "ga", "gm", "ge", "de", "gh", "gi", "gr", "gl", "gd", "gp", "gu", "gt", "gn", "gw", "gy", "ht", "hm", "va", "hn", "hu", "is", "in", "id", "ir", "iq", "ie", "il", "it", "jm", "jp", "jo", "kz", "ke", "ki", "kp", "kr", "kw", "kg", "la", "lv", "lb", "ls", "lr", "ly", "li", "lt", "lu", "mo", "mk", "mg", "mw", "my", "mv", "ml", "mt", "mh", "mq", "mr", "mu", "yt", "mx", "fm", "md", "mc", "mn", "me", "ms", "ma", "mz", "mm", "na", "nr", "np", "nl", "an", "nc", "nz", "ni", "ne", "ng", "nu", "nf", "mp", "no", "om", "pk", "pw", "ps", "pa", "pg", "py", "pe", "ph", "pn", "pl", "pt", "pr", "qa", "re", "ro", "ru", "rw", "sh", "kn", "lc", "pm", "vc", "ws", "sm", "st", "sa", "sn", "rs", "sc", "sl", "sg", "sk", "si", "sb", "so", "za", "gs", "ss", "es", "lk", "sd", "sr", "sj", "sz", "se", "ch", "sy", "tw", "tj", "tz", "th", "tl", "tg", "tk", "to", "tt", "tn", "tr", "tm", "tc", "tv", "ug", "ua", "ae", "gb", "um", "uy", "uz", "vu", "ve", "vn", "vg", "vi", "wf", "eh", "ye", "zm", "zw".

Optional. Added to the location.

Optional. Added to the location.

Optional. Added to the location.

Optional. Added to the location.

A GeoJSON geometry. Its centre is used as the map centre for the search. The outline itself is not applied, so results near the edge are not clipped.

Case-insensitive match against a business’s categories. Filtered-out businesses are not written and not charged.

"all" keeps everything Google returned. "only_includes" keeps names containing the search term. "only_exact" keeps exact name matches. Default: "all". Accepted: "all", "only_includes", "only_exact".

Drop businesses rated below this. Businesses with no rating are dropped when a minimum is set. Default: "". Accepted: "", "two", "twoAndHalf", "three", "threeAndHalf", "four", "fourAndHalf".

Keep only businesses that have a website, or only those without one. Default: "allPlaces". Accepted: "allPlaces", "withWebsite", "withoutWebsite".

Turn off the anonymous reliability telemetry described in the README. When this is on, nothing about your run — including your search terms — leaves the run. Default: false.

Google's feature ID, e.g. 0x8644b4eb0b01986f:0x6b241d4769ed491e. The stable key for joining and de-duplication.

Google Place ID, e.g. ChIJ…

Customer ID, the decimal form used by maps?cid= links.

Knowledge Graph machine ID, e.g. /g/11abc…

Google Business Profile ID.

Google Maps link for this business.

Business name.

Primary category, the first entry of categories.

All categories Google lists, primary first.

Editorial summary. Null when Google has none.

Short tagline. Additional field: the reference Actor does not publish this.

Street address without the business name, e.g. 1700 S 1st St, Austin, TX 78704.

street

city

state

postalCode

Two-letter country code.

neighborhood

{ lat, lng }, each null when Google has no coordinates.

Phone as Google displays it.

Phone in E.164-like digits.

website

Average rating, 1 to 5.

Number of reviews. Null when this response variant omitted it.

Weekly hours as { day, hours }, Monday first. Empty when the business publishes none.

True when all seven days are present. Google returns a short variant at random; the run retries, but not every business ends up complete.

Open or closed text at the moment of the scrape, e.g. "Closed · Opens 9 AM Sun". NOT a permanent-closure flag.

IANA timezone, e.g. America/Chicago.

Accessibility attributes as { group: [{ label: boolean }] }. The search response carries the accessibility group only.

Preview image URL.

The search term this row came from.

Language used for the search.

Position within that search term's results, starting at 1.

ISO timestamp of when this row was produced.

Version 2 is the default anonymous feed. Original publication time and edit status are unknown (null). Choose 1 for the legacy cookie feed. A run cannot resume a Dataset or STATE created with the other version. Default: 2. Minimum: 1. Maximum: 2.

One place per line. Four forms are accepted and all resolve to the same place: a Google Maps link (https://www.google.com/maps?cid=… or /maps/place/…), a ChIJ… place ID, a numeric CID, or a 0x…:0x… feature ID. Anything else is skipped, reported in the run summary, and never charged.

All four are Google's own orders, not a re-sort on our side. Newest follows Google and orders by last edit time, which is what updatedTime carries. Default: "mostRelevant". Accepted: "mostRelevant", "newest", "highest", "lowest".

Upper bound per place. A place with fewer reviews returns what it has. Version 2 fetches up to 60 reviews per page; version 1 fetches 20. Default: 100. Minimum: 1. Maximum: 100000.

BCP-47 code such as en, de, ja, pt-BR. This is Google's interface language: it changes the relative time wording ("3 weeks ago" vs "vor 3 Wochen") and the language Google would translate into. It does NOT filter reviews — review text always comes back in the language the reviewer wrote it in. Anything unrecognised falls back to en. Default: "en".

Legacy contract version 1 only. Google sometimes answers an anonymous session with a short five-review response that has no next page, which looks exactly like a genuine last page. On: retry with fresh sessions and, if the response is still short, mark every row incomplete: true. Off: accept the short response silently. Default: true.

How many places to scrape at the same time. Delivery order is preserved regardless. Default: 2. Minimum: 1. Maximum: 8.

Legacy contract version 1 only. How many fresh sessions to try when Google returns a truncated response, before the Actor re-queues the whole place and starts over after a longer pause. The truncation is bound to the session, and swapping sessions within a couple of seconds tends to land on the same one, so a low number here plus the re-queue is both faster and more reliable than a high number. Default: 3. Minimum: 0. Maximum: 8.

Legacy contract version 1 only. Pause before opening a fresh session. Retrying immediately lands on the same truncated variant more often. Default: 1500. Minimum: 0. Maximum: 10000.

Legacy contract version 1 only. Experimental opt-in. Reuses short-lived Google Maps sessions across Data Worker instances. Enable only for controlled latency and completeness comparisons. Default: false.

Google's own review id. Stable across sessions and sort orders, so it is the key to de-duplicate and join on.

The review in the language the reviewer wrote it in. Not a machine translation.

Language of text, as Google reports it.

1 to 5.

Contract 2: null because the original publication time is unavailable. Contract 1: original publication time in ISO 8601 when available.

Observed update timestamp in ISO 8601, or null when unavailable. Contract 2 cannot determine original publication time or edit status. Contract 1 falls back to time for unedited reviews. Google's newest order follows this field.

Contract 2: null because edit status is unknown. Contract 1: whether time and updatedTime differ.

Google's own wording, e.g. "2 months ago".

Public reviewer identity as Google shows it on the review. Nothing inferred.

The business reply, when there is one. Only a relative time is available.

Photo URLs attached to the review.

Which place this review belongs to, plus the input it came from.

Which order this run requested.

page

Position within this place, in the requested order.

Contract 2: full identifies the anonymous feed and does not guarantee completeness. Contract 1: full or teaser identifies the legacy response variant. See incomplete for known delivery limits.

True when we know this place was not fully delivered. See the README on truncated responses.

expected

scrapedAt

contractVersion

Google interface language this run asked for. Review text is always the original language; this only affects relative time wording.

One domain per line, for example acme-widgets.de. A full URL works too; www. is stripped and duplicates are removed.

Website URLs to start from, when you want a specific page instead of the homepage. Combined with the domains above.

Hard cap on how many domains are written to the dataset across the whole run, including a resumed run. Minimum: 1. Maximum: 1000000.

Homepage plus up to four more: contact, about, team and legal/imprint. Lower it to 1 to scan homepages only — that is faster and cheaper, and in our own 1,039-domain test it found addresses on 39.9% of domains instead of 53.7%. Default: 5. Minimum: 1. Maximum: 5.

Keep gmail.com, outlook.com and similar addresses when they are published as the site's contact address. Small businesses often use one. They are always labelled separately from addresses on the site's own domain. Default: true.

Adds mxAvailable to each address: whether the email domain publishes MX records. This is a DNS fact about the domain, not a check of the individual mailbox and not a deliverability guarantee. Default: true.

How many batches of 8 domains run at once. Raise it for large lists, lower it if you want a gentler crawl. Default: 4. Minimum: 1. Maximum: 10.

The input domain, lower-cased and without www.

Exactly what you passed in.

The URL that was actually reached, after redirects.

Only contacts_found is charged. Accepted: "contacts_found", "no_contacts_found", "unreachable".

Fetch result of the homepage: ok, http_403, http_429, dns_failed, timeout, tls_failed, challenged, http_5xx, …

emailCount Minimum: 0.

Addresses on the input domain, subdomains included. Minimum: 0.

gmail, outlook and similar, published as the site contact address. Minimum: 0.

otherDomainCount Minimum: 0.

Website-builder and tooling addresses that were dropped. Minimum: 0.

One entry per address, each with its source URL, page type and surrounding text.

Every page that was fetched and how it ended.

requests Minimum: 0.

tookMs Minimum: 0.

contractVersion Minimum: 1.

One destination per line, for example "Lisbon, Portugal". A comma is handled for you: Airbnb geocodes "Paris, France" to a different city unless the separator is normalised, which this Actor does.

Airbnb room links (airbnb.com/rooms/…) or plain numeric room IDs. Order is preserved. Invalid entries are skipped and never charged.

Off: search results only, charged as listing-delivered. On: one extra public request per listing for amenities, house rules, sub-ratings, host and images, charged as detailed-listing-delivered. Direct room URLs always fetch details. Default: false.

Airbnb caps a single search at 15 pages. Pages overlap, so a destination yields roughly 220 unique listings at most. Default: 100. Minimum: 1. Maximum: 450.

YYYY-MM-DD. Give both dates or neither. Without dates Airbnb picks a stay window per listing, so prices are not comparable across rows.

YYYY-MM-DD.

Number of adults for the stay. Affects which listings Airbnb returns and how they are priced. Default: 1. Minimum: 1. Maximum: 16.

Number of children aged 2 to 17. Default: 0. Minimum: 0. Maximum: 15.

Number of infants under 2. Airbnb does not count them toward a listing capacity. Default: 0. Minimum: 0. Maximum: 5.

Number of pets. Listings that do not allow pets are filtered out by Airbnb. Default: 0. Minimum: 0. Maximum: 5.

Prices come back in this currency, as Airbnb formats them. Default: "USD".

Language of names, descriptions and amenity labels. Default: "en-US".

Airbnb filters on the nightly price. The price field on each row is the stay total, so a row's own number will not fall inside this range. Minimum: 0. Maximum: 100000.

Airbnb filters on the nightly price. The price field on each row is the stay total, so a row own number will not fall inside this range. Minimum: 0. Maximum: 100000.

One extra public request per host for about text, hosting years, host rating and review count, listing count, verification and profile highlights such as languages. Charged at the same rate; a failed enrichment never fails the listing. Default: false.

Airbnb room ID. The stable key for joining and de-duplication.

Room URL with the stay parameters used.

Listing name.

Airbnb's type-and-place line, e.g. "Apartment in Lisbon".

subtitle

"search" when the row came from a destination search, "direct_url" when it came from a room URL.

"search_result" when a price is present. Null for direct-URL rows: Airbnb does not expose booking prices to anonymous listing pages.

{ label, price, qualifier, originalPrice } as Airbnb formats them. Null for direct-URL rows.

The window Airbnb priced, e.g. "Nov 29 – Dec 4".

checkIn

checkOut

Null in this version: Airbnb does not expose booking availability on anonymous listing pages.

rating

reviewsCount

Rating as shown, e.g. "4.81 (209)". "New" for new listings.

Six sub-ratings plus overall and review count. Only when details were fetched.

{ latitude, longitude }.

propertyType

roomType

personCapacity

homeTier

{ title, items } such as "4 guests", "1 bedroom".

Bed and bath line from the search card.

description

htmlDescription

descriptionOriginalLanguage

seoTitle

metaDescription

Neighbourhood sections. Titles are available; the body text is not exposed to anonymous pages, so content is usually null.

breadcrumbs

Grouped amenities, including entries marked unavailable.

houseRules

highlights

De-duplicated across search card and detail page.

Such as "Guest favorite".

Such as "Free cancellation".

Identity always; about, hosting years, host rating and listing count only when host enrichment ran.

coHosts

androidLink

iosLink

Whether the detail page was fetched for this row.

Whether host enrichment succeeded.

Set when the detail fetch failed and the row fell back to search data.

hostError

The destination this row came from. Null for direct-URL rows.

page

rank

scrapedAt

contractVersion

Websites to analyse. Accepts example.com, www.example.com or https://example.com/path — the host is extracted for you. Entries that cannot be a domain are skipped and never charged. Default: [].

Upper bound on unique domains processed in one run. Duplicates collapse to a single lookup and a single result. Default: 1000. Minimum: 1. Maximum: 5000.

How many batches of 16 domains run in parallel. The upstream rate limit is the real ceiling, so raising this mostly adds waiting. Default: 2. Minimum: 1. Maximum: 10.

Normalized host that was looked up.

Site title. Empty string when the source has none.

Site description. Empty string when the source has none.

Two-level category label, e.g. finance/banking_credit_and_lending.

Global traffic rank. Null for sites without one, including many subdomains and link shorteners.

Rank in the site's top country: Country, CountryCode, Rank. All null when absent.

Rank inside the category, as a string.

Visits in the snapshot month, as a string. "0" is a real zero.

Bounce rate from 0 to 1, as a string.

Average pages per visit, as a string.

Average seconds on site, as a string.

Three months of visits keyed by YYYY-MM-01.

Share of visits by source, each 0 to 1: SocialOrganic, SocialPaid, Mail, Referrals, SearchOrganic, SearchPaid, Direct, GenAi, Affiliate, DisplayAds.

Visit share by country: Country, CountryCode, Value.

Top organic keywords: name, esitmatedValue, cpc. The misspelled key is the source's and is preserved.

Empty array. The public source used here does not return competitor lists.

True when the site owner publishes its own analytics.

Source-hosted preview image URL.

Month the numbers describe.

The domain exactly as the caller typed it.

The cleaned form that was queried. Internationalized names keep their Unicode spelling and are never converted to punycode.

Whether the source has measurable traffic for this site. False rows are written so the list stays aligned and are never charged.

Organic search volume per keyword, aligned with topKeywords: name, volume.

AI referral traffic: available, totalVisits, share (0 to 1), distributionBoundary, sources (per LLM with rank, share and monthlyHistory) and prompts. Unknown numbers are null; a confirmed absence is an empty array.

"ads" returns individual ad creatives. "advertisers" returns one summary row per advertiser — a cheap way to check which companies run Google Ads. Default: "ads". Accepted: "ads", "advertisers".

Free-text brand names (e.g. "Nike") or bare domains (e.g. "nike.com"). Brand names are resolved to advertisers through the Transparency Center's own autocomplete; anything that looks like a domain is used directly.

Exact domains to pull ads for. Returns every advertiser account whose ads point at that domain.

Specific advertiser IDs starting with "AR", as they appear in an advertiser's Transparency Center URL.

Paste Transparency Center URLs: advertiser pages, creative pages, or domain search pages. They are parsed automatically.

ISO 3166-1 alpha-2 country code (US, DE, GB …), or ANYWHERE for no country filter. Several codes can be given comma-separated; results are the union. Default: "US".

Restrict to ads served on specific Google surfaces. Leave empty for all. Default: [].

Filter by creative format. Several formats can be given comma-separated. Default: "ALL". Accepted: "ALL", "TEXT", "IMAGE", "VIDEO".

Only ads shown on or after this date (YYYY-MM-DD).

Only ads shown on or before this date (YYYY-MM-DD).

Maximum ad creatives to collect for each domain or advertiser. Default: 100. Minimum: 1. Maximum: 50000.

Hard cap on delivered rows across all targets. This is what you pay for. Default: 1000. Minimum: 1. Maximum: 100000.

When a brand-name query resolves to several advertisers, how many of the top matches to scrape. Default: 5. Minimum: 1. Maximum: 50.

On by default. Adds every creative variation, the per-country reach with each country's last-shown date, and EU impression ranges. Turn it off for a faster, list-only run. Default: true.

How many detail batches to fetch in parallel. Default: 2. Minimum: 1. Maximum: 8.

Turn off the anonymous reliability telemetry described in the README. When this is on, nothing about your run — including your search terms — leaves the run. Default: false.

Transparency Center creative ID, e.g. CR09405416287280037889.

Transparency Center advertiser ID, e.g. AR16735076323512287233.

Advertiser display name as Google publishes it.

Landing domain the ad points at. Null when the creative was reached by advertiser or creative URL, where the search row is not available.

TEXT, IMAGE or VIDEO. Unknown enum values come through as the raw number.

ISO timestamp of the first recorded impression.

ISO timestamp of the most recent recorded impression.

Days the creative has been running, as reported by Google. Null when absent; never 0 for unknown.

"image" for an archived bitmap preview with a direct URL, "iframe" for a Google-hosted render page. Null when neither is present.

Direct CDN URL of the archived image preview. Null for render-page previews.

Google-hosted preview render page. Not a raw asset URL. Null for archived image previews.

Pixel width of the archived image preview. Null for render-page previews.

Pixel height of the archived image preview. Null for render-page previews.

Google's internal creative ID, parsed from the preview URL. Null for plain image ads.

Obfuscated Google Ads customer ID, parsed from the preview URL. Groups ads from the same ad account. Null for plain image ads.

Ad group ID parsed from the preview URL. Clusters ads from the same campaign. Null for plain image ads.

Creative version ID parsed from the preview URL. Null when absent.

Transparency Center link to this creative.

The country filter this row was collected under, echoed back. ANYWHERE when no country filter was used.

Number of creative variations. Only present when includeDetails is on.

Every creative variation: type, imageUrl, previewUrl, width, height and internal IDs. Only present when includeDetails is on.

Every country the ad ran in, each with that country's last-shown date. Only present when includeDetails is on.

EU DSA transparency block: available, impressions.lowerBound / upperBound, firstShownDate, lastShownDate and platformDistribution. available is false when Google publishes none for this creative. Only present when includeDetails is on.

Advertiser display name.

ISO 3166-1 alpha-2 country the advertiser is registered in.

Lower bound of Google's approximate ad count for this advertiser. Null when Google publishes no range.

Upper bound of Google's approximate ad count. Null when Google publishes no range.

The input that produced this row.

Transparency Center link to the advertiser page.

LinkedIn Ad Library search URLs or LinkedIn company URLs. Search parameters such as accountOwner, payer, keyword, countries, dates, impressions, targeting, and sort are preserved.

Maximum successfully delivered ads per input URL. Leave empty to continue until LinkedIn reports the true end or your run budget is reached. Minimum: 1. Maximum: 100000.

Fast list mode. Returns the public search-card fields, which may contain truncated text and fewer details. Full details are recommended. Default: false.

Turn off the anonymous reliability telemetry described in the README. When this is on, nothing about your run — including the keywords in your Ad Library URLs — leaves the run. Default: false.

LinkedIn Ad Library ad identifier.

Canonical public Ad Library detail URL.

Public advertiser logo or profile image URL.

Advertiser name displayed by LinkedIn.

Public LinkedIn company or member URL.

First and last shown dates when LinkedIn exposes them.

Full public ad copy in detail mode; possibly truncated in list mode.

Public creative or CTA destination URL.

Public call-to-action labels.

Observed LinkedIn creative format.

Creative headline when publicly present.

Primary image creative URL.

Highest-bitrate public MP4 URL.

Public document PDF URL.

All public document page images.

Message Ad sender image.

Entity LinkedIn identifies as paying for the ad.

Public impression range.

Country impression percentages when published.

Top public targeting parameters.

Thought Leader ad member name.

Thought Leader public member URL.

Thought Leader member image.

Thought Leader public headline.

Original input source URL that discovered the ad.

Keywords to search Pinterest for. Each keyword is paginated until the run hits Max results.

Paste Pinterest URLs: a pin, a profile, a profile's Created or Saved tab, a board, or a search results page. Each is routed automatically.

What keyword searches return. Pins and videos come back as pin records; boards and profiles come back as board and profile records. Default: "all-pins". Accepted: "all-pins", "videos", "boards", "profiles".

Hard cap on delivered rows across every keyword and URL. This is what you pay for. Default: 100. Minimum: 1. Maximum: 100000.

On by default. Adds one lookup per pin for video files, merchant attribution and product metadata. Turn it off for a faster, list-only run. Default: true.

Optional. Paste the lastBookmark from a previous run's output to continue paginating the first target where it stopped.

How many detail batches to fetch in parallel. Default: 2. Minimum: 1. Maximum: 8.

"pin", "video", "board" or "profile".

The record's entity: pin, board or profile. Videos are pins.

Contract version and entity type.

Pinterest ID of the pin, board or profile.

Canonical Pinterest URL for the record.

Pin title, board name or profile display name.

The Pinterest page this record came from.

"query" when the record came from a keyword, "url" when it came from a pasted URL.

The keyword or URL that produced this record.

Pin fields: title, description, alt_text, created_at, link, domain, dominant_color, is_promoted, is_video, reaction_counts, story, product_metadata_type, rich_summary_type. Present on pin records.

The pin's creator: id, username, full_name, follower_count, three avatar sizes, profile url, merchant and identity verification flags.

The board the pin sits on: id, name, url, pin_count, section_count, collaborator_count, is_collaborative, cover images and the board owner.

images (thumb / small / medium / large / original plus the raw images_by_size table) and video (hls_url, mp4_url, duration_ms, width, height, thumbnail, variants). video is null for still pins.

Merchant attribution block, present on shopping pins once details are fetched.

Whether this row went through the per-pin detail lookup. Tells apart "the source has no value" from "we did not look".

Board record fields, including description and follower_count. Present on board records.

Profile record fields, including about, pin_count, board_count, following_count and website_url. Present on profile records.