Trade Press Scraper

Turn advertising and marketing trade press into structured data. Pick a date range, optionally narrow it with a Boolean query, and get one row per article with the headline, publish time, author, section, tags, summary, campaign credits where the publication prints them, and every video, social post and outbound link the article points to.

Covers two publications, in one dataset with the same columns:

• Marketing-Interactive (marketing-interactive.com): APAC marketing and advertising news for Singapore, Malaysia, Hong Kong, Indonesia, the Philippines, Thailand and Australia, archive back to 2007.
• Campaign Brief Asia (campaignbriefasia.com): campaign launches and agency news across Asia and the Middle East, archive back to 2019. About half its articles end with a credits block (client, agency, production house, media agency, director), which comes back as separate columns.

What you can use it for

• New business and competitive tracking. Agency appointments, account wins, pitches and reviews for the brands and agencies you care about.
• Campaign libraries. Every campaign a brand or agency ran in a period, with links to the films on YouTube or Vimeo.
• Who made what. Which agencies, production houses and media agencies a brand worked with, from the published campaign credits.
• Market and media monitoring. Mentions of brands, agencies, people or topics across APAC trade press, on a schedule.
• Research. How coverage of a topic (retail media, AI, sustainability) moved over months or years.

How it works

1. Choose the publications and a date range ("Published from" accepts a date or a range like 2 years).
2. Optionally write a Boolean query. Matching against headline and tags checks every article before downloading it, so only matches are fetched. (Campaign Brief Asia has no tags, so there the headline is checked.) Matching against full article text downloads everything in the range and checks the body too, which catches names that only appear in campaign credits.
3. Run it. Each article becomes one row. Download as JSON, CSV or Excel, or pull it through the Apify API.

Writing a query

NOT binds tightest, then AND, then OR, so a OR b AND c means a OR (b AND c). Use brackets when in doubt. Operators must be uppercase.

Example queries:

(appoint* OR pitch* OR "account" OR review*) AND (Ogilvy OR Publicis OR Dentsu)

(Grab OR Gojek OR foodpanda) AND (campaign* OR film OR launch*)

(CPF OR HPB OR "Health Promotion Board" OR govt OR ministr*)
AND NOT (Malaysia OR Indonesia)

Each row's matched_terms lists which parts of the query it matched, so you can see why an article was included. A query with a syntax error stops the run before anything is fetched, with a message saying what is wrong.

Tip: headlines often use informal names ("govt", "Health ministry") and lowercase acronyms ("MoH"). Include the variants, or switch to full-text matching.

Input

Article text

By default each row carries the article's metadata, credits and links, and body_text is empty. The body is still read, so word counts, credits, links, videos and full-text query matching all work.

The publication's own boilerplate inside the article (related headlines, "See more" links, event ads) is removed from the text, so a full-text query doesn't match an article because of a headline it merely links to.

Turn on Include article text to receive the full text as well. Article text belongs to its publisher: use it for your own research and analysis, and don't republish it.

Output

One row per article. A real row from Marketing-Interactive:

{
  "site": "marketing-interactive",
  "url": "https://www.marketing-interactive.com/cpf-gets-weird-to-show-retirement-planning-can-be-easier",
  "title": "CPF gets weird to show retirement planning can be easier",
  "published_at": "2026-10-01T09:21:03+08:00",
  "author": "A'bidah Zaid Shirbeeni",
  "author_url": "https://www.marketing-interactive.com/authors/a-bidah-zaid-shirbeeni",
  "section": "News",
  "tags": ["CPF Board", "MoneySense", "GOVT VCCP", "retirement planning",
           "CPF campaign", "Starcom Worldwide", "Salihin Ramli", "Gabe Chen"],
  "teaser": "CPF Board and MoneySense are taking a lighter approach to retirement planning, using fortune tellers, fortune cats and an overload of financial information to show Singaporeans that preparing for the future does not have to be complicated.",
  "word_count": 534,
  "video_urls": [
    "https://www.youtube.com/watch?v=QQ16ojXd3GY",
    "https://www.youtube.com/watch?v=dDSGLut-VD8",
    "https://www.youtube.com/watch?v=EvYVOA9Jl8k"
  ],
  "social_urls": [],
  "external_links": [],
  "internal_links": [
    "https://www.marketing-interactive.com/cpf-concludes-integrated-marketing-pitch",
    "https://www.marketing-interactive.com/how-public-sector-campaigns-in-sg-have-become-increasingly-un-government"
  ],
  "image_url": "https://marketing-interactive-assets.b-cdn.net/article_images/cpf-gets-weird-to-show-retirement-planning-can-be-easier/1790817651_CPF%20%281%29.jpg",
  "matched_terms": [],
  "body_text": null,
  "scraped_at": "2026-10-08T12:26:18.995Z",
  "error": null
}

(Tags and internal links shortened here.)

A real row from Campaign Brief Asia, showing the credit fields:

{
  "site": "campaign-brief-asia",
  "url": "https://campaignbriefasia.com/2026/10/07/ocbc-and-govt-vccp-singapore-show-how-ai-can-make-anyone-look-like-a-scammer/",
  "title": "OCBC and GOVT VCCP Singapore show how AI can make anyone look like a scammer",
  "published_at": "2026-10-07T08:28:39+08:00",
  "author": "Adam Shaw",
  "section": "Agencies",
  "credit_client": "OCBC",
  "credit_agency": "GOVT VCCP",
  "credit_production": "ROLLA Productions",
  "credit_media": "Wavemaker",
  "credits": [
    {"group": null, "role": "Client", "name": "OCBC"},
    {"group": null, "role": "Agency", "name": "GOVT VCCP"},
    {"group": null, "role": "Media Partner", "name": "Wavemaker"},
    {"group": null, "role": "Film Director", "name": "Jake Nam"},
    {"group": null, "role": "Film Production", "name": "ROLLA Productions"}
  ],
  "video_urls": ["https://www.youtube.com/watch?v=-c9b1LjobcM"]
}

(Other fields left out here.)

Rows with error set keep the same columns, so exports line up. Filter them out by keeping rows where error is empty.

How credits are read

Credits come only from the article's labelled credits block, never guessed from the text. A line such as Production House: Freeflow Productions fills credit_production; lines naming people (Director: Peggy Goh,

Head of Production: ...

How long does it take?

Pages are fetched at 20 per minute by default, which keeps the run gentle on the publication and clear of its rate limits.

• A week of both publications (about 175 articles): about 9 minutes.
• Two years filtered by headline and tags: usually a few hundred articles, under half an hour.
• Two years unfiltered: around 9 hours for Marketing-Interactive (about 10,300 articles), 3 hours for Campaign Brief Asia (about 3,700). Narrow the query or the date range, or raise the speed towards 60 per minute.

For runs longer than an hour, check the run's Timeout in the run options. A run that reaches its timeout stops and keeps what it collected so far.

Pricing

Pay per event. See the Pricing tab for current prices.

• Article: each article in your results.
• Article text: an add-on per article, only when Include article text is on.
• Article scanned: only with full-text matching, for each article that was downloaded and checked against your query but didn't match. Same price as an article, so every article read costs the same whether it matches or not.
• Actor start: once per run.

Pages that fail are never charged. With headline-and-tags matching, non-matching articles are never downloaded, so they cost nothing.

As a guide, at the time of writing:

• A week of both publications (about 175 articles): about $1.60.
• Two years of Marketing-Interactive (about 10,300 articles): about $93, or about $196 with article text on every row.
• Two years of Campaign Brief Asia (about 3,700 articles): about $34. Five years (about 8,600): about $78.
• A full-text search reads every article in the range, so it costs the same as pulling them all, plus article text for the matches if you turn it on.
• A headline-and-tags search only pays for its matches: one matching 400 articles over two years costs about $4.

Set a maximum cost per run in Apify and the run stops when it is reached. Each run's final status lists what was charged.

FAQ

Does it get around paywalls? No. It reads publicly available pages only, and only publications whose articles are free to read are supported.

Is scraping these sites allowed? The actor reads public pages, follows each publication's robots.txt, and is rate-limited. You are responsible for how you use the data, including the publisher's terms and copyright. If in doubt, take legal advice.

Why did my query return fewer articles than expected? With headline-and-tags matching, an article only counts if your terms appear in its headline or tags. Add informal variants, or switch to full-text matching.

Why are the credit columns empty on some rows? Either the article has no credits block (about half of Campaign Brief Asia articles don't, and Marketing-Interactive doesn't use one), or no line in it carries that role. Check credits for the full block.

Why are some rows errors? A page that stays unavailable after several retries becomes an error row with the reason, so nothing goes missing silently. Rows saying the site is throttling mean the speed was too high: lower Max pages per minute.

Can I schedule it? Yes. Save a task with a relative date such as 7 days and schedule it weekly in Apify.

Limitations

• Two publications so far: Marketing-Interactive and Campaign Brief Asia.
• Credit columns are filled only from a labelled credits block. Marketing- Interactive writes credits into the article prose, so finding them there needs full-text matching.
• The headline used for headline matching is the one in the publication's sitemap or feed, which very occasionally differs slightly from the one on the page.
• Speed is capped at 60 pages a minute to protect the publication.