# MGID Ads Scraper - Native Ads by Country, Payer (`automation_craft/mgid-ads-scraper`) Actor

Scrape the native ads MGID runs right now in the sponsored boxes of 914 verified publisher sites, read from 48 countries, or any page URL. One row per distinct ad: headline, image, brand, ad id, call to action, category, and the EU payer and beneficiary MGID discloses. Watch mode bills only new ads.

- **URL**: https://apify.com/automation_craft/mgid-ads-scraper.md
- **Developed by:** [Automation Craft](https://apify.com/automation_craft) (community)
- **Categories:** Marketing, Lead generation, SEO tools
- **Stats:** 3 total users, 2 monthly users, 100.0% runs succeeded, 0 bookmarks
- **User rating**: No ratings yet

## Pricing

from $0.85 / 1,000 ads

This Actor is paid per event. You are not charged for the Apify platform usage, but only a fixed price for specific events.
Since this Actor supports Apify Store discounts, the price gets lower the higher subscription plan you have.

Learn more: https://docs.apify.com/actors/running/actors-in-store.md#pay-per-event

## What's an Apify Actor?

An Actor is a serverless cloud program that runs on the Apify platform. It has two run modes.
In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours,
and optionally produces a well-defined JSON output, datasets with results, or files in key-value store.
In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server.

Apify vocabulary and the platform model are defined once, in the agent quickstart at https://apify.com/agents.md.

## How to integrate an Actor?

If asked about integration, you help developers integrate Actors into their projects.
You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready.

Do not guess an integration path. Every one of them is in the agent quickstart at https://apify.com/agents.md: the Apify MCP server, Agent Skills with the Apify CLI, the JavaScript and Python clients, the REST API, and the account-free path for an agent with no human to sign in. It also carries the rule on stating cost before the first paid run.

For examples already wired to this Actor's own input schema, see the [API](#api) section below.

Each client library has reference documentation the quickstart does not restate: [JavaScript/TypeScript](https://docs.apify.com/api/client/js/docs.md) (`npm install apify-client`) and [Python](https://docs.apify.com/api/client/python/docs.md) (`pip install apify-client`).

# README

### MGID Ads Scraper - Native Ads by Country, Payer

**MGID ads scraper** for the native ads running right now in MGID's sponsored content boxes on news and publisher sites. Pick countries, pick sites, or paste any article URL, and get one row per distinct ad: headline, image, brand, MGID's ad id, call to action, category number, and for readers in the EU and the UK the **payer, on whose behalf the ad runs and the payer's country**, as MGID discloses them. No login, no API key, no browser. Each distinct ad is charged once per run, and a watch mode charges only ads your earlier runs have not seen.

This is a live sampler of the sponsored content MGID serves today, per country and per site, built for affiliate marketers, media buyers and agencies who study competitor native ads, and for compliance teams who want the payer behind a native ad. It is not an ad library: MGID publishes no archive, so there is no advertiser search and no history before your own runs (see "What this Actor does NOT do").

#### Quick start

1. Open the Actor and keep the prefilled input (United States, 3 verified sites, 20 ads), or pick your countries.
2. Optionally pick named sites from the built in list, or paste article URLs of any site that shows an MGID box.
3. Click Start. A run of the prefilled input takes about 10 to 30 seconds.
4. Read the **Ads** view: its first column, `type`, tells the rows apart (`ad` for an ad of an advertiser, charged; `house_ad` for MGID's own content, free; the run's coverage and summary rows follow with empty cells, since a dataset view cannot filter by row type). The **Coverage** view has one row per site and country: calls, answers with and without ads, distinct ads.
5. To track new native ads every day, give the run a **memory name** and schedule it: each later run delivers and charges only ads the memory has not seen.

### What you get

Each ad row carries the fields below. Fill rates are measured on 519 distinct ads delivered by this Actor's own test runs on 2026-10-11 (24 runs on the final build; United States, Germany, Brazil, France, United Kingdom, Italy and Armenia exits; 130 of them read from EU and UK exits).

| Field | What it is | Filled |
|---|---|---|
| `headline` | Headline: the ad headline as the box shows it, HTML entities decoded | 100 percent |
| `brandName` | Brand: the brand name shown with the ad | 100 percent |
| `adId` | Ad id: MGID's id of the ad, the same on every site and call | 100 percent |
| `imageUrl` | Image: the ad image as MGID serves it | 100 percent |
| `callToAction` | Call to action: the button text, in the reader's language (Learn more, Mehr erfahren) | 100 percent |
| `categoryId` | Category number: MGID's category of the ad (MGID publishes no name table for it) | 100 percent |
| `mediaType` | Media type: `static` for a still image, `anim` for an animated one | 100 percent (anim on 2.5 percent) |
| `payerName`, `onBehalfOf`, `payerCountry` | Who pays for the ad, on whose behalf it runs and the advertiser's country, as MGID discloses them to readers in the EU and the UK (read through MGID's own payer information, one small call per new ad) | 80 percent of ads read from EU and UK exits (104 of 130); none elsewhere |
| `description` | Description: a second line of text, when the ad has one | 1.3 percent |
| `aiGenerated` | AI generated: true when MGID marks the ad as made with AI | true on 0.8 percent |
| `landingUrl`, `landingDomain` | Landing address, only when MGID states one in its answer (rare; MGID's answer carries the click counter, never followed and never delivered) | 0 of 519 in these runs (5 of 853 in the build's spike) |

Also on every ad row: `payerShown` (`shown`, `not_shown` when the box shows no payer, which is the case for most readers outside the EU and the UK, or `lookup_failed`), `personalDataMasked`, the `site`, its own `siteCountry`, the `widgetId`, `pageUrl` and `exitCountry` it was seen on, its `position` in the box, `seenAt`, `creativeId` and `adKey`.

The run also writes free rows:

- **house_ad**, MGID's own content in the box (brands such as Brainberries, Herbeauty and Buzz Day, all paid for by MGID's own company as MGID's payer data says): the same fields as an ad row plus `houseReason` (`house_payer` when the payer data says so, `house_brand` for a brand known as MGID's own). Delivered once per run, never charged, never counted against `maxAds`; turn them off with `includeHouseAds: false`.
- **coverage**, one per site and country read: calls, answers with ads, answers without ads (MGID serves only its own content to callers it does not trust, so such an answer is retried on a fresh exit and never read as "this site shows no ads"), distinct ads, ads delivered, ads already seen on another site in the run, ads your memory already knew, MGID house rows seen, payer lookups, their keys, where the widget id came from, and why the site stopped (`dry`, `site_call_cap`, `blocked`, `failed`, `unknown_widget`, `bad_widget`, `no_mgid_box`, `stopped`).
- **status**, for a value that could not be used (a page URL that is not a URL or is not accepted), a country with no verified site, or the pairs a stopped run did not reach.
- **summary**, the last row: calls, page reads, payer lookups, ads seen and delivered, ads charged, house rows, the stop reason, proxy bytes and the work allowance.

#### Built in sites

914 publisher sites in 46 countries are built in. They come from MGID's own list of its publishers (its sellers.json file): each site's home page was read once for its MGID box (the widget on the page, or the widget list of MGID's site loader the page names), and each site was then verified live on 2026-10-11: three calls on three fresh residential exits in the site's country, at least two of them answering with ads of advertisers. The country of a site is read from its domain ending, else from the page language, and a hand review moved 126 sites to their own country (English pages without a regional mark are listed under the United States). Pick them by country (the sites with the most ads per call first) or by name. The list is refreshed by a script (the site map), so sites that drop MGID leave it. Australia and Canada have no MGID site of their own in the map; they are offered because MGID answered ads of advertisers to readers there on other countries' sites (measured 2026-10-11), and they are read on the best other verified sites.

| Country | Sites | Sites, the most ads per call first (all of them are in the Sites field of the input) |
|---|---|---|
| United States (`US`) | 177 | fingerlickingrecipes.com, gazettely.com, policefilestonite.net, freepatternsblog.com, abdullah-al-sahil.com, aic5.com, bullscore.com, dailyusanews.online, daytondailymagazine.com, detroitcitynews.com, emedihealth.com, epochedge.com, gamingnews4u.com, happiness-life.org, healthtips8386.blogspot.com, htdmedia168.blogspot.com, idropnews.com, instanthub.net, isolastories.com, mydailystars.com, patriotclash.com, patriotsgazette.com, rawlifestories.site, realtwistedstories.com, salondiscover.com and 152 more |
| United Kingdom (`GB`) | 6 | footballtoday.com, en31.usnews.uk, lolitopia.com, operatorsd.com, rungbeg.com, thecityisours.co.uk |
| Germany (`DE`) | 17 | ddr-rezepte.net, donnerwetter.de, lilienblog.de, uelzener-nachrichten.com, quermania.de, marktcom.de, finanznachrichten.de, muligartenbaugmbh.de, reitschuster.de, juraforum.de, wallstreet-online.de, boersennews.de, ideecon.com, sharedeals.de, pravda-tv.com, besserleben.blog, globalimagine.com |
| France (`FR`) | 9 | astucesjardin.com, thebuzzly.com, amigo-royan.fr, adoboloco.fr, brasseriemoliteuil.fr, meilleurpronostic.fr, emet.press, gnatepe.com, scolinfo.net |
| Italy (`IT`) | 33 | aforismi.meglio.it, calabriadasogno.it, jedanews.com, pescarasport24.it, sicurauto.it, tg24-ore.com, worldnotix.net, epgitalia.tv, ilblogdigio.it, ligurianotizie.it, ideadesigncasa.org, spazionapoli.it, cultweb.it, professionedocente.com, sfornopizza.it, nuovasocieta.it, fineartphotographyvideoart.com, ketosisa.it, zmedia.it, stiledonna.net, billoccino.com, gds.it, lasmorfianapoletana.com, phonetoday.it, ristorantevangoghmatera.it and 8 more |
| Spain (`ES`) | 15 | skynovels.net, aquienlasierra.es, somosfan.com, notienred.info, noticiasprofeticas.com, codigooculto.com, 180grados.digital, noticiaalminuto.com, aqueenofmagic.com, chicanoticias.com, depeso.com, elfiltrador.com, laisladelastentaciones.com, ncrnoticias.com, todosobreremedios.com |
| Netherlands (`NL`) | 1 | dissident.one |
| India (`IN`) | 42 | indianews.in, braveindianews.com, janamtv.com, vikatan.com, mediaoneonline.com, wikiwiki.in, aajkaal.in, janmabhumi.in, ibtimes.co.in, talkindiadigital.com, ukhrultimes.com, lokshahi.news, taasir.com, thefederal.com, bigbreakingwire.in, dainiksaveratimes.com, rajyasameeksha.com, uttranews.com, mycgnews.in, bartamanpatrika.com, saamana.com, hindi.webdunia.com, page16.in, aninews.in, theinsiderjournal.in and 17 more |
| Australia (`AU`) | 0 | no site of its own: read on other countries' verified sites from its own exit (topUpWithOtherSites) |
| Canada (`CA`) | 0 | no site of its own: read on other countries' verified sites from its own exit (topUpWithOtherSites) |
| Albania (`AL`) | 4 | sot.com.al, sarandanews.net, agroweb.org, tetovasot.com |
| Algeria (`DZ`) | 1 | lematindalgerie.com |
| Argentina (`AR`) | 9 | canal26.com, chacodiapordia.com, entreriosya.com.ar, historiahoy.com.ar, perfil.com, elobservadornews.com.ar, elpatagonico.com, infobrisas.com, paparazzi.com.ar |
| Armenia (`AM`) | 1 | news.am |
| Azerbaijan (`AZ`) | 1 | sonxeber.az |
| Bosnia and Herzegovina (`BA`) | 2 | bih-zmajevi.com, kakanj-x.com |
| Brazil (`BR`) | 170 | bemparana.com.br, agoranovale.com.br, diariodorio.com, artesanatopassoapassoja.com.br, bocanews.com.br, clickpetroleoegas.com.br, gmaisnoticias.com, informenoticias.net, jasb.com.br, livrodosonho.com, mizaeltutodroid.com, mrnews.com.br, portaledicase.com, portalumuaramanews.com.br, receitatodahora.com.br, revistaartesanato.com.br, sociedademilitar.com.br, termometrodapolitica.com.br, teuquiz.com, carroesporteclube.com.br, adonisdutra.com.br, bahiainforma24horas.com, bhaz.com.br, fanoticias.com.br, financenews.com.br and 145 more |
| Bulgaria (`BG`) | 3 | receptite.com, kalvacha.com, kliuki.bg |
| Chile (`CL`) | 3 | diarioelcentro.cl, latribuna.cl, eldesconcierto.cl |
| Colombia (`CO`) | 14 | laopinion.co, hoydiariodelmagdalena.com.co, vibra.co, lanacion.com.co, elpilon.com.co, noticiasuno.com, winsports.co, elquindiano.com, telemedellin.tv, pulzo.com, teleamiga.tv, elespectador.com, eluniversal.com.co, quindionoticias.com |
| Costa Rica (`CR`) | 1 | lateja.cr |
| Croatia (`HR`) | 1 | epodravina.hr |
| Cyprus (`CY`) | 1 | omegalive.com.cy |
| Czech Republic (`CZ`) | 5 | morezprav.cz, psiusmev.cz, cestazaremeslem.cz, stars24.cz, novinykraje.cz |
| Denmark (`DK`) | 3 | tipsbladet.dk, campo.dk, mandesiden.dk |
| Estonia (`EE`) | 1 | avasta.me |
| Ghana (`GH`) | 1 | ghanablogs.com |
| Greece (`GR`) | 28 | araxthos.gr, financenews.gr, england365.gr, metrosport.gr, pao365.gr, dokari.gr, evianews.com, eviathema.gr, grand-prix.gr, fanpage.gr, flight.com.gr, paidis.com, dailypost.gr, apopsilive.gr, europost.gr, i-diakopes.gr, sofokleousin.gr, exposgreece.gr, monobala.gr, live-sports365.com, parapolitika.gr, tvxs.gr, enimerotiko.gr, menshouse.gr, triklopodia.gr and 3 more |
| Guatemala (`GT`) | 2 | deguate.com, prensalibre.com |
| Honduras (`HN`) | 1 | televicentro.com |
| Hungary (`HU`) | 5 | magyarhaza.com, tudasfaja.com, mosolyogjvelunk.com, kulfoldiapro.com, balramagyar.hu |
| Indonesia (`ID`) | 120 | jpnn.com, ayojakarta.com, burungnews.com, gosipgarut.id, harga.web.id, jawapos.com, beritajatim.com, inews.id, manadopost.jawapos.com, airspace-review.com, arif.rahmawan.web.id, batarapos.com, beritainspiratif.com, beritasampit.com, beritasatu.com, explorethebali.com, goriau.com, guesehat.com, harga-diskon.com, hariankepri.com, indomiliter.com, indosport.com, industry.co.id, inisumedang.com, intrik.id and 95 more |
| Kenya (`KE`) | 5 | afrinewske.com, businesstoday.co.ke, kenyanlife.com, kenyanlife.info, malindikenya.net |
| Malaysia (`MY`) | 8 | sinarharian.com.my, thesun.my, defencesecurityasia.com, flashsukan.com.my, kosmo.com.my, sabahpost.net, thecoverage.my, portalsemakan.com |
| Mexico (`MX`) | 23 | xeu.mx, tusbuenasnoticias.com, vecinosinformados.com, am.com.mx, criteriohidalgo.com, capitalmexico.com.mx, culturacolectiva.com, diario.mx, elimparcial.com, laopcion.com.mx, amexi.com.mx, marthadebayle.com, record.com.mx, merca20.com, unavida.website, imparcialoaxaca.mx, reflectores.mx, abcnoticias.mx, tribunadelabahia.com.mx, gluc.mx, dariocelis.mx, semanario7dias.com.mx, quien.com |
| Nigeria (`NG`) | 9 | naijanews.com, abntv.com.ng, emmanuelsblog.com.ng, fusionchronicles.ng, blueprint.ng, gistlover.com, nairametrics.com, thenewsguru.ng, gwg.ng |
| North Macedonia (`MK`) | 6 | ubavinaizdravje.mk, opserver.mk, kurir.mk, mastilo.mk, ekipa.mk, nmd.mk |
| Peru (`PE`) | 8 | walac.pe, larotativa.pe, elpopular.pe, notitas.site, moda.com.pe, tvolima.pe, larepublica.pe, pachamamaradio.org |
| Philippines (`PH`) | 2 | governmentph.com, rmn.ph |
| Poland (`PL`) | 29 | jebzmeme.pl, infor.pl, udostepnijto.pl, dsi.net.pl, burbonik.pl, doba.pl, film.org.pl, tkn24.pl, tuwroclaw.com, infoilawa.pl, learningfromhollywood.pl, tysol.pl, chords.pl, crowdmedia.pl, efakty24.pl, eszamotuly.pl, miejskireporter.pl, pogonsportnet.pl, silesia24.pl, superprzepis.pl, tuolawa.pl, wlocie.pl, zdroweporadniki.pl, zlubaczowa.pl, zwierzaki.pl and 4 more |
| Portugal (`PT`) | 14 | postal.pt, imprensa.com.pt, noticiasdecoimbra.pt, bolanarede.pt, diarioluso.com, revistacarros.pt, buzzvip.pt, rumores.pt, leonino.pt, flash.pt, blogvisaodemercado.pt, portalfresta.com, sulinformacao.pt, bix.pt |
| Romania (`RO`) | 6 | dej24.ro, adevarul.ro, ziarulincomod.ro, mytex.ro, 101books.club, rotalianul.com |
| Serbia (`RS`) | 4 | sanovnik.org, glas-javnosti.rs, trebinjelive.info, hellomagazin.rs |
| South Africa (`ZA`) | 1 | news365.co.za |
| Thailand (`TH`) | 11 | brighttv.co.th, tvpoolonline.com, dog-vs-cat.com, esanborleumtin.com, js100.com, pattayamail.com, thaistar24h.net, thainewspost.com, innnews.co.th, banmuang.co.th, mthai24h.com |
| Turkey (`TR`) | 9 | gazetepencere.com, haberlisin.com, ozgurhalkinsesi.com.tr, veryansintv.com, kamuajans.com, olaymedya.com, haberdarim.com, basvuruformu.org, memurlar.net |
| Ukraine (`UA`) | 69 | bbc-ccnn.com, gromada.group, volyn.com.ua, shostka.info, ua.news, dozi-rok.com, itechua.com, bukinfo.com.ua, novosti-n.org, pik.net.ua, racurs.ua, rivne.media, trueua.info, firtka.if.ua, golapravda.blog, fraza.com, izvestia.kiev.ua, cikavosti.com, goloskarpat.info, insider.ua, sq.com.ua, borgexpert.com, myzukrainy.net, politerno.com.ua, rivnepost.rv.ua and 44 more |
| Vietnam (`VN`) | 33 | movieshd4k.com, hot.newsonline.biz, baoangiang.com.vn, baodongnai.com.vn, vietnamfinance.vn, thethao247.vn, phapluat.suckhoedoisong.vn, thoitiet.vn, ixinh.com.vn, kienthuc.net.vn, vietnamplus.vn, ghienbongda.vn, suckhoedoisong.vn, ifeg.info, onsports.vn, vietstock.vn, vov.vn, baomoi.vip, tinnhanh247.online, toploigiai.vn, nguoi-noi-tieng.com, diembao.org, anninhthudo.vn, doctor247.vn, giaoducthoidai.vn and 8 more |

Any other site that shows an MGID box can be read with **Page URLs**: paste one article URL per site. The Actor reads that page once to find the site's MGID widget id (from the widget on the page, or from MGID's site loader the page names), then calls the box like a built in site. A page without an MGID box gets a free `no_mgid_box` coverage row; a widget MGID no longer knows gets a free `unknown_widget` row.

#### How a run reads the boxes

- Every call goes out through a fresh Apify residential exit in the country asked for. MGID serves the ads of the reader's country, not of the site (three American sites read from German exits and from US exits showed no ad in common), so a built in site is read from its own country, page URLs from every country you pick, and a thin country is topped up with other verified sites read from its own exit (`topUpWithOtherSites`; the rows name the site's own country in `siteCountry`).
- MGID writes the reader's time zone offset and whether it applied the GDPR to the reader into its answer. An answer whose offset or GDPR flag does not fit the exit country (a proxy exit placed in another country) is not delivered and counts as a failed call. An exit placed in another country with the same offset and the same GDPR status (Germany and France, for example) cannot be told apart this way.
- MGID rotates the ads it shows: each call to a site can bring new ones, up to the size of the site's box (1 to 32 ads). A site is called again until it stops bringing ads new to the run (`stopAfterDryCalls`), up to `maxCallsPerSite`.
- Unpaid calls are capped too: answers without new ads (a watch run whose memory already knows most ads, blocked or empty answers, page reads of sites that are not built in) draw on a small budget per run that grows with every ad charged. When it runs out the run stops with `stoppedBy: work_allowance` in the summary; the sites it had not finished get a coverage row with status `stopped` and the same `stoppedBy`, and sites never reached are listed in one `not_started` status row. Start the next run (or schedule it) to go on; ads already delivered are not charged again under a memory name.
- The Actor reads the same data call the publisher's page makes and MGID's own payer information for an ad, and stops there. It never clicks an ad, never follows a click link and never loads tracking pixels: a followed link would be a paid click on an advertiser's campaign.

### How much does it cost to scrape MGID ads?

Pay per event: a small start fee per run and a price per ad delivered; never a charge for calls, page reads, payer lookups, MGID's own content, blocked answers, coverage rows or status rows.

| Event | Free plan | Bronze | Silver | Gold |
|---|---|---|---|---|
| Ad (per 1,000 distinct ads) | USD 1.45 | USD 1.25 | USD 1.05 | USD 0.85 |
| Actor start (once per run) | USD 0.008 | USD 0.008 | USD 0.008 | USD 0.008 |

Platinum and Diamond plans pay the Gold price. The Actor start event is the same on every plan.

The Store pricing card shows these same prices per 1,000 events: "$1.45 / 1,000" on the Ad row means one ad costs 0.145 cents.

Worked example on the Free plan: the prefilled run (20 ads) costs USD 0.008 + 20 x USD 0.00145 = USD 0.037. A run of 500 ads costs USD 0.008 + USD 0.725 = USD 0.733.

- An ad is charged once per run, however many sites, calls or countries show it. MGID writes the reader's city into many headlines ("Wichita Men: ..."); the same ad with another city in its headline is the same ad (the row keeps the first headline seen).
- With a memory name, an ad is charged once for that name: later runs deliver only new ads.
- MGID's own content (house rows), answers without ads, page reads, payer lookups, unknown widgets, failed calls, coverage rows, status rows and the summary are free.
- Set a maximum cost per run in the run options: the Actor stops before an ad it cannot charge, and never makes a call when the limit cannot pay for one more ad.

### Input

| Field | What it does | Default |
|---|---|---|
| `countries` | Countries to sample; verified sites of each are read from a residential exit in that country | United States, when no site and no page URL is given |
| `sitesPerCountry` | How many verified sites of each chosen country to read, the most ads per call first (0 reads none by country) | 5 |
| `topUpWithOtherSites` | Top up a country with fewer verified sites of its own than `sitesPerCountry` with other verified sites, read from that country's exit | true |
| `sites` | Named sites from the built in list, each read from its own country | none |
| `pageUrls` | Any article or page URL of any site, read from every chosen country (the United States when no country is chosen) | none |
| `maxAds` | The most ads delivered in one run (house rows are not counted) | 100 |
| `maxCallsPerSite` | The most calls to one site in one run | 10 |
| `stopAfterDryCalls` | A site stops after this many answers in a row with no ad new to the run | 3 |
| `maxCalls` | The most requests in one run (page reads of sites that are not built in included; payer lookups are not counted) | 200 |
| `includeHouseAds` | Deliver MGID's own content as free `house_ad` rows | true |
| `memoryName` | Watch mode: one name for scheduled runs, each ad charged once for that name | none |
| `resetMemory` | Forget what the named memory knows before this run | false |
| `includePersonalData` | Personal data: on, `brandName`, `payerName` and `onBehalfOf` are delivered exactly as MGID serves them, e-mail addresses included; off, an e-mail address in them is hidden | false |

Example: every new native ad in Germany and France with its payer, daily.

```json
{
    "countries": ["DE", "FR"],
    "sitesPerCountry": 10,
    "maxAds": 1000,
    "memoryName": "de-fr-daily"
}
```

The `sites` field takes only names from the built in list: a site that is not in it (also one written with other capitals) is refused by Apify before the run starts, with an error that lists the allowed values; read any other site with `pageUrls`. A value that cannot be used is never guessed: a page URL that is not a URL, an address on an ad network's own host, or a URL that carries an ad click id, an e-mail address or a password gets a free status row and the rest of the run goes on; an unknown country or a malformed number is refused before any call is made, and Apify itself refuses a `true` or `false` setting written as text before the run starts.

### What this Actor does NOT do

- It is not an ad library. MGID publishes no ad archive, so there is no search by advertiser or keyword, no history before your own runs and no "all ads of brand X". Filter the rows by `brandName` or `payerName` for a competitor's ads, and use a memory name to build your own history.
- No landing pages: MGID's answer carries a click counter, not the advertiser's address (a landing address is stated on under 1 percent of ads, and delivered then). The Actor never follows a click link or resolves a redirect (a followed link is a paid click).
- No advertiser id, campaign id, spend, impressions, clicks or first and last shown dates: MGID's answer carries none of them. No active status or date filters either: every row is an ad being served at the moment it was seen.
- No lookup by advertiser domain or brand and no media download or rehosting: the image addresses are MGID's own, and filtering by `brandName` or `payerName` is done on the rows.
- The payer is shown only where MGID discloses it, for readers in the EU and partly in the UK; ads read from other countries have no payer. Payer names can be a company or a person's name and are delivered as MGID publishes them.
- MGID's own content is recognised by its payer (where MGID shows one) and by the brands that payer is known to pay for; a new brand of MGID's own, seen only where no payer is shown, would be delivered as an ad.
- It does not cover every MGID site: the built in list holds the sites verified live, and page URLs reach any other one. A box that shows no ads of advertisers three times in a row on fresh exits is reported as `blocked`.
- It does not read Outbrain, Taboola or other native networks (Outbrain and Taboola have their own listings of the same family).
- It does not sign in or keep cookies: every call is a fresh visitor with no history in the chosen country, so the sample is what a new anonymous reader there is shown.
- It does not use any publisher API key, and needs no MGID account.

### FAQ

#### Is there a public ad library for MGID?

No. MGID publishes no ad archive. This Actor reads the ads MGID serves right now in the boxes of publisher sites, per country, and keeps what your runs saw in a named memory if you ask it to.

#### How do I spy on competitors' MGID native ads?

Run the countries where they advertise with enough sites and calls, then filter the rows by `brandName` or, for EU and UK readers, by `payerName` and `onBehalfOf`. Schedule the run with a memory name to collect their new ads every day.

#### Who pays for an MGID ad?

For readers in the EU (and partly the UK), MGID discloses the payer, on whose behalf the ad runs and the advertiser's country next to each ad. The Actor reads that disclosure for every new ad and puts it into `payerName`, `onBehalfOf` and `payerCountry`.

#### Which countries and sites are covered?

914 publisher sites in 46 countries are built in, each verified live from a residential exit in the site's country (United States 177, United Kingdom 6, Germany 17, France 9, Italy 33, Spain 15, Netherlands 1, India 42, and more; the full list is in "Built in sites" above). A country with fewer sites of its own than you ask for is topped up with other verified sites read from that country's exit (MGID serves the ads of the country the reader is in, so a Brazilian site read from a German exit shows the ads MGID serves to readers in Germany). Any other site that shows an MGID box can be read with a page URL, from any of the 48 offered countries. Australia and Canada are offered with no site of their own and read on other countries' sites.

#### Does the Actor click ads or visit landing pages?

No. It reads the same data call the publisher's page makes for its box, and MGID's own payer information for an ad, and stops there: no click link is followed, no landing page is loaded and no tracking pixel is sent. For a page URL of a site that is not built in, it reads that page once, and MGID's site loader file when the page names one, to find the box.

#### Does the Actor deliver personal data?

A payer or a brand can be a person, and a name can hold an e-mail address. By default an e-mail address in `brandName`, `payerName` or `onBehalfOf` is replaced by "(e-mail address hidden)" and `personalDataMasked` is true on that row; a person's name without an e-mail address is delivered as MGID publishes it. Turn on `includePersonalData` to get those three fields exactly as MGID serves them. You are responsible for a lawful use of personal data, GDPR included.

#### Why does this Actor run with limited permissions?

Least privilege: it reads and writes only its own run storages and, in watch mode, the named memory store it creates itself. It cannot see your other actors, datasets or account data.

### API examples

```bash
curl -X POST "https://api.apify.com/v2/acts/automation_craft~mgid-ads-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"countries": ["US"], "sitesPerCountry": 3, "maxAds": 50}'
```

```javascript
import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: 'YOUR_TOKEN' });
const run = await client.actor('automation_craft/mgid-ads-scraper').call({ countries: ['DE'], sitesPerCountry: 5, maxAds: 200 });
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items.filter((i) => i.type === 'ad').map((i) => [i.brandName, i.payerName, i.headline]));
```

```python
from apify_client import ApifyClient
client = ApifyClient("YOUR_TOKEN")
run = client.actor("automation_craft/mgid-ads-scraper").call(run_input={"countries": ["FR"], "maxAds": 200, "memoryName": "fr-daily"})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    if item["type"] == "ad":
        print(item["brandName"], item["payerName"], item["headline"])
```

### Changelog

- 0.1 (2026-10-19): new setting `includePersonalData`, off by default; on, the brand, payer and beneficiary names are delivered exactly as MGID serves them, e-mail addresses included. Runs without it are unchanged.
- 0.1 (2026-10-11): first version. MGID native ads by country, built in site or page URL; the EU payer from MGID's own disclosure; MGID's own content as free house rows; one row and one charge per distinct ad per run; watch mode with a named memory; free coverage rows per site and country.

### More data tools by Automation Craft

- [Outbrain Ads Scraper - Native Ads, Advertisers](https://apify.com/automation_craft/outbrain-ads-scraper)
- [Taboola Ads Scraper - Native Ads by Country](https://apify.com/automation_craft/taboola-ads-scraper)
- [Meta Ads Library Scraper - Ad Library, All Filters](https://apify.com/automation_craft/meta-ads-library-scraper)
- [Google Ads Transparency Scraper - Decoded Creatives](https://apify.com/automation_craft/google-ads-transparency-scraper)
- [TikTok Ads Library Scraper: EU Ads, No Login](https://apify.com/automation_craft/tiktok-ads-library-scraper)
- [LinkedIn Ad Library Scraper: Ads by Company](https://apify.com/automation_craft/linkedin-ad-library-scraper)
- [Microsoft Ads Library Scraper - Competitor Ads](https://apify.com/automation_craft/microsoft-ads-library-scraper)
- [Amazon Sponsored Products Scraper - Ads by Keyword](https://apify.com/automation_craft/amazon-sponsored-products-scraper)

# Changelog

This Actor's version history is a separate document: https://apify.com/automation_craft/mgid-ads-scraper/changelog.md

# Actor input Schema

## `countries` (type: `array`):

Countries to sample. For each, verified MGID publisher sites of that country are read from a residential exit in it (MGID serves the ads of the reader's country); 46 countries have verified sites of their own, and Australia and Canada are read on other countries' verified sites from their own exits (see Top up thin countries). Leave empty with no sites and no page URLs to read the United States.

## `sitesPerCountry` (type: `integer`):

How many of each chosen country's verified sites to read, the sites with the most ads per call first. 0 reads no site by country (only the sites and page URLs below). Default 5.

## `topUpWithOtherSites` (type: `boolean`):

When a chosen country has fewer verified sites of its own than Sites per country, read the best other verified sites from that country's exit to make up the number (MGID serves the ads of the reader's country on every site; each row names the site and its own country). Default true.

## `sites` (type: `array`):

Verified publisher sites to read, each from a residential exit in the site's country (914 sites). Adds to the sites picked by country.

## `pageUrls` (type: `array`):

Any article or page URL of any site that shows an MGID box, read from every chosen country (the United States when no country is chosen). For a site that is not built in, the Actor reads the page once to find its MGID widget id; a page without an MGID box gets a free coverage row.

## `maxAds` (type: `integer`):

The most ads delivered in one run (each distinct ad is one row and one charge). Default 100.

## `maxCallsPerSite` (type: `integer`):

The most times one site is read in one run, each time from a fresh exit (MGID rotates the ads it shows, so more calls find more ads). Default 10.

## `stopAfterDryCalls` (type: `integer`):

A site stops when this many answers in a row brought no ad new to the run. Default 3.

## `maxCalls` (type: `integer`):

The most requests in one run, over all sites and countries (page reads for sites that are not built in included; MGID's payer information calls are not counted). Default 200.

## `includeHouseAds` (type: `boolean`):

MGID fills part of its boxes with its own content (brands paid for by MGID's own company). When on, each such item is delivered once per run as a free house_ad row; it is never charged. Default true.

## `memoryName` (type: `string`):

Give scheduled runs one name and each ad is delivered and charged once for that name: later runs deliver only ads the memory has not seen. Ads the memory knows are counted in the coverage rows, free. Letters, digits, spaces and . \_ , - : # ( ), up to 100 characters.

## `resetMemory` (type: `boolean`):

Forget every ad the named memory knows before this run, so every ad counts as new again.

## `includePersonalData` (type: `boolean`):

Off by default. Off, an e-mail address in the brand (brandName), the payer (payerName) or on whose behalf the ad runs (onBehalfOf) is replaced by "(e-mail address hidden)" and the row's personalDataMasked is true, as in every earlier version. On, those three fields are delivered exactly as MGID serves them, e-mail addresses included. You are responsible for a lawful use of personal data, GDPR included.

## Actor input object example

```json
{
  "countries": [
    "US"
  ],
  "sitesPerCountry": 3,
  "topUpWithOtherSites": true,
  "maxAds": 20,
  "maxCallsPerSite": 4,
  "includeHouseAds": true,
  "includePersonalData": false
}
```

# Actor output Schema

## `items` (type: `string`):

No description

# API

You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup.

## JavaScript example

```javascript
import { ApifyClient } from 'apify-client';

// Initialize the ApifyClient with your Apify API token
// Replace the '<YOUR_API_TOKEN>' with your token
const client = new ApifyClient({
    token: '<YOUR_API_TOKEN>',
});

// Prepare Actor input
const input = {
    "countries": [
        "US"
    ],
    "sitesPerCountry": 3,
    "topUpWithOtherSites": true,
    "maxAds": 20,
    "maxCallsPerSite": 4,
    "includeHouseAds": true
};

// Run the Actor and wait for it to finish
const run = await client.actor("automation_craft/mgid-ads-scraper").call(input);

// Fetch and print Actor results from the run's dataset (if any)
console.log('Results from dataset');
console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`);
const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((item) => {
    console.dir(item);
});

// 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs

```

## Python example

```python
from apify_client import ApifyClient

# Initialize the ApifyClient with your Apify API token
# Replace '<YOUR_API_TOKEN>' with your token.
client = ApifyClient("<YOUR_API_TOKEN>")

# Prepare the Actor input
run_input = {
    "countries": ["US"],
    "sitesPerCountry": 3,
    "topUpWithOtherSites": True,
    "maxAds": 20,
    "maxCallsPerSite": 4,
    "includeHouseAds": True,
}

# Run the Actor and wait for it to finish
run = client.actor("automation_craft/mgid-ads-scraper").call(run_input=run_input)

# Fetch and print Actor results from the run's dataset (if there are any)
print(f"💾 Check your data here: https://console.apify.com/storage/datasets/{run.default_dataset_id}")
for item in client.dataset(run.default_dataset_id).iterate_items():
    print(item)

# 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start

```

## CLI example

```bash
echo '{
  "countries": [
    "US"
  ],
  "sitesPerCountry": 3,
  "topUpWithOtherSites": true,
  "maxAds": 20,
  "maxCallsPerSite": 4,
  "includeHouseAds": true
}' |
apify call automation_craft/mgid-ads-scraper --silent --output-dataset

```

## MCP server setup

```json
{
    "mcpServers": {
        "apify": {
            "type": "http",
            "url": "https://mcp.apify.com/?tools=fetch-actor-details,automation_craft/mgid-ads-scraper"
        }
    }
}
```

The hosted server signs you in with OAuth on first connect, so no API token belongs in this config. Clients without OAuth support can send an `Authorization: Bearer <APIFY_API_TOKEN>` header instead, using a token from API & Integrations in Apify Console (https://console.apify.com/settings/integrations).

## OpenAPI specification

Download the OpenAPI definition: https://api.apify.com/v2/actors/ixNJtfoEiU5kuVnmr/builds/VpYUkSfjN9FgGhIja/openapi.json