Detik News Scraper

Scrapes news articles from Detik.com, Indonesia's largest online news network. For each article it collects the headline, summary, author, publish and update timestamps, section, keywords, images and the full article body. Results are returned as a structured JSON dataset that can be exported to CSV or Excel.

Target website: detik.com

💡 Use Case

Media monitoring teams, researchers and analysts who need a continuously updated, machine-readable feed of Indonesian news coverage without copying articles by hand.

🏭 Industry

Media & Entertainment, Market Research, Marketing & Advertising, Academic Research.

📤 Output

Structured JSON dataset (exportable to CSV/Excel) with one record per article, containing: url, title, description, authors, publisher, publishedAt, modifiedAt, category, section, keywords, images, body and wordCount.

🌐 Domain

Online news and media content extraction (Indonesia).

🏷️ Label

detik, news-scraper, indonesia, article-scraper, media-monitoring, content-extraction

✨ Features

• Article extraction — headline, summary, author, publish and update timestamps, section/category, keywords, images and the full article text from Detik.
• Pagination support — walks listing pages automatically until the per-URL item limit is reached.
• Proxy support — optional Apify proxy configuration (datacenter or residential).
• Automatic URL detection — listing and article URLs are told apart automatically, so you can mix both in Start URLs.
• Whole-site mode — one toggle crawls every major section instead of a hand-picked list.
• Cross-run caching — an optional project name skips articles already scraped in an earlier run.
• AI enrichment — optionally paraphrases each article section by section, writes a short summary, and analyses its sentiment, using any of several models through OpenRouter.

🔧 Input Configuration

Field	Type	Required	Default	Description
startUrls	array	✅	[{"url": "https://finance.detik.com/indeks"}]	Detik listing URLs (channel front pages, /indeks pages or search results) or direct article URLs. The type of each URL is detected automatically.
maxItems	integer	—	5	Maximum number of articles to scrape for each start URL.
scrapeWholeSite	boolean	—	false	Ignore Start URLs and crawl the index pages of every major Detik channel instead.
cacheProjectName	string	—	—	Optional. When set, articles already scraped under this project name in a previous run are skipped instead of being scraped again.
aiEnabled	boolean	—	false	Use an AI model to paraphrase each article, summarize it, and analyse its sentiment. Runs on the Actor's own AI credits — no API key needed from you.
aiModels	array	—	["openai/gpt-4o-mini"]	Models tried in order — the first that returns a usable result wins, so the rest act as fallbacks.
aiFeatures	array	—	["summarize", "paraphrase", "sentiment"]	Which enrichments to request: summarize, paraphrase, sentiment, keywords, custom.
aiCustomInstructions	string	—	—	Extra analysis to run, used only when custom is among the selected AI features.
proxyConfiguration	object	—	{"useApifyProxy": false}	Detik responds fine without a proxy. Enable datacenter proxies first if you start seeing blocks, and residential only if datacenter still fails.

Supported URL types:

• Listing pages — channel index pages such as https://finance.detik.com/indeks, channel front pages such as https://news.detik.com/, and search result pages such as https://www.detik.com/search/searchall?query=ekonomi.
• Article pages — any URL containing a /d-{id}/ segment, e.g. https://finance.detik.com/bursa-dan-valas/d-8656440/dolar-as-menguat-tipis-ke-rp-17-522.

📦 Output

The dataset has one view, Overview, a compact table of each article's title, category, authors, publish date, word count and URL. The full record contains every field listed above.

Sample output:

{
    "url": "https://finance.detik.com/berita-ekonomi-bisnis/d-8656439/bulog-kawal-penyaluran-bantuan-pangan-beras-di-cirebon-agar-tepat-sasaran",
    "title": "BULOG Kawal Penyaluran Bantuan Pangan Beras di Cirebon agar Tepat Sasaran",
    "description": "BULOG mengawasi penyaluran Bantuan Pangan Beras di Cirebon, memastikan distribusi tepat sasaran. Tanggapi dugaan penyalahgunaan dengan menghormati proses hukum.",
    "authors": ["Inkana Putri"],
    "publisher": "detikcom",
    "publishedAt": "2026-09-10T09:20:03+07:00",
    "modifiedAt": "2026-09-10T09:20:08+07:00",
    "category": "Ekonomi Bisnis",
    "section": "detikFinance",
    "keywords": [
        "bnr",
        "bulog",
        "bantuan pangan",
        "ketahanan pangan",
        "distribusi bantuan",
        "cirebon",
        "pemerintah desa",
        "bhabinkamtibnas",
        "desa ambit",
        "indonesia",
        "perum bulog jawa barat",
        "babinsa",
        "nurman susilo",
        "kabupaten cirebon",
        "kecamatan waled",
        "tni",
        "polri",
        "tim",
        "kawal",
        "pbp",
        "penyaluran",
        "sasaran",
        "papar",
        "penerima",
        "kelancaran",
        "bulog cabang cirebon",
        "bulog kawal penyaluran bantuan pangan beras",
        "karung beras",
        "pangan beras",
        "bulog jawa",
        "pemimpin wilayah perum bulog jawa barat",
        "bulog cabang",
        "bulog kawal",
        "penyaluran bantuan pangan beras"
    ],
    "images": [
        "https://awsimages.detik.net.id/community/media/visual/2026/09/10/bulog-kawal-penyaluran-bantuan-pangan-beras-di-cirebon-pastikan-tepat-sasaran-1789006704387_169.jpeg?w=1200"
    ],
    "body": "Perum BULOG terus mengawal penyaluran Bantuan Pangan Beras agar manfaatnya diterima oleh masyarakat yang berhak. Melalui satuan kerja di Cabang Cirebon, BULOG memastikan setiap tahapan distribusi dimonitor secara optimal dalam mendukung program pemerintah untuk menjaga ketahanan pangan sekaligus memenuhi kebutuhan masyarakat.\n\nMenanggapi informasi dugaan penyalahgunaan 198 karung beras Bantuan Pan …",
    "wordCount": 325
}

🤖 AI enrichment output

When aiEnabled is on, each article gains an ai object. Requesting several features costs a single model call per article, so the record shape stays the same whichever features you pick.

Field	Type	Description
ai.summary	string	A 2–3 sentence summary of the article.
ai.paraphrase	array of strings	The article rewritten section by section, in the original order — one array element per section. Kept roughly the same length as the source, so it is a paraphrase and not a summary.
ai.sentiment	string	"positive", "neutral" or "negative".
ai.keywords	array of strings	5–10 key phrases. Only when keywords is selected.
ai.custom	string	The result of aiCustomInstructions. Only when custom is selected.
ai.model	string	Which model actually produced the result — useful when a fallback was used.

Unselected features come back as null. If every model fails, ai is null and the article is still saved with all its scraped fields, so enrichment problems never cost you the crawl.

Pricing. Enrichment is billed as a pay-per-event charge of $0.05 per enriched article, on top of the usual platform usage. You are charged once per article that comes back enriched — never per model, so a retry on a fallback model costs the same as a first-try success, and an article no model could enrich is not charged at all. Leaving aiEnabled off incurs no AI charges whatsoever.

Sample enrichment — real output from the article Masih Kesulitan Bikin Laporan Keuangan Usaha? Coba Belajar Accurate (297 words), enriched by openai/gpt-4o-mini:

"ai": {
    "summary": "Mengelola keuangan bisnis bisa menjadi tantangan, terlebih bagi UMKM dan fresh graduate yang tidak berpengalaman dengan perangkat lunak akuntansi. Accurate + Sertifikasi CAP #13 menawarkan kursus online untuk membantu peserta memahami perangkat akuntansi, dengan kesempatan mendapatkan sertifikat setelah mengikuti pelatihan. Program ini ditujukan untuk pelaku UMKM, mahasiswa akuntansi, dan profesional yang ingin meningkatkan keterampilan akuntansi digital mereka.",
    "paraphrase": [
      "Mengatur keuangan dalam bisnis bukan hal yang selalu sederhana, terutama bagi pelaku usaha mikro, kecil, dan menengah (UMKM) yang masih menggunakan metode pencatatan manual. Saat volume transaksi meningkat, mengidentifikasi pendapatan, biaya, dan kinerja laba bisa menjadi tantangan yang cukup rumit.",
      "Hal ini juga berlaku bagi fresh graduate atau mahasiswa di bidang akuntansi. Walaupun teori akuntansi telah dipelajari di perguruan tinggi, tidak selalu berarti mereka sudah mahir dalam menggunakan aplikasi akuntansi yang umumnya dipakai dalam lingkungan kerja. Ketika mereka memasuki dunia profesional dan diminta untuk mengoperasikan perangkat seperti Accurate, pengalaman praktik menjadi nilai tambah yang sangat berharga.",
      "Oleh sebab itu, keterampilan dalam menggunakan perangkat lunak akuntansi digital tidak hanya penting bagi staf perusahaan. Pebisnis UMKM yang ingin mengatur keuangannya secara lebih terstruktur, serta fresh graduate yang ingin mempersiapkan diri sebelum memulai karier, juga dapat meraih keuntungan dari kemampuan ini.",
      "Permintaan untuk pelatihan ini menjadi inti dari Accurate + Sertifikasi CAP #13, yang diselenggarakan oleh Piranha Smart Center (PSC) melalui Education Visionary Lab (EVL).",
      "… 5 more sections, in the article's original order"
    ],
    "keywords": null,
    "sentiment": "positive",
    "custom": null,
    "model": "openai/gpt-4o-mini"
  }

The paraphrase is returned as an ordered array, one element per section of the source article, so you can walk it in the same sequence the story was written in. Together its 9 sections run to about the same length as the original — it rewrites the article rather than shortening it.

💡 Uses of This Data

• Tracking how a topic, company or public figure is covered over time
• Building a searchable archive of Indonesian news coverage
• Training or evaluating Indonesian-language NLP models
• Competitive and brand monitoring across news channels
• Feeding dashboards or newsletters with fresh headlines
• Sentiment tracking across Indonesian coverage of a brand, sector or public figure
• Republishing paraphrased coverage where quoting the original verbatim is not an option

⚠️ Limitations & Known Issues

• Live front pages change constantly — running the same listing URL twice may return a different set of articles as new stories are published.
• AI enrichment is billed per article and slows the crawl — $0.05 per enriched article, plus roughly 4–5 seconds per article on top of scraping. It is off by default.
• AI output is model-generated — summaries, paraphrases and sentiment labels should be reviewed before being published or used for decisions.
• Very long articles are truncated for enrichment — only the first 8,000 characters are sent to the model. The scraped body field is always complete.
• Section pages list a fixed window — a listing page only exposes the stories it currently shows; use /indeks pages with pagination to reach further back.

📝 Notes

• No proxy needed by default — Detik responds reliably without one, which keeps run costs down. Enable datacenter proxies first if you start seeing blocks, and residential only if datacenter still fails.
• API integration — this actor can be called as an API from any automation platform (Zapier, Make, n8n, custom scripts).

🚀 How to Use

1. Sign up for a free Apify account — includes $5 monthly credit.
2. Open the actor page and click Try for free.
3. Fill in the Input fields (at least one start URL is required).
4. Click Start and wait for the run to complete.
5. Download results from the Output tab in JSON, CSV, or Excel format.

You can also run this actor via the Apify API or integrate it directly into your workflows using Zapier, Make, or n8n.

⚖️ Legality of this actor

This actor only collects data that is already publicly visible on Detik.com — it does not log in, bypass a paywall, or access private content. Scraping publicly available data is generally considered lawful (see hiQ Labs v. LinkedIn as precedent). You remain responsible for complying with Detik's Terms of Service and with any applicable laws such as GDPR if you process personal data contained in article text or bylines.

🔗 Other Actors

• 🚗 Cheapest Carsandbids Scraper — Use this cheap actor to scrape auction listings, sale prices, vehicle specs, and photos from Cars & Bids [carsandbids.com] website.
• 🌐 WHOIS Domain Lookup Scraper — Bulk WHOIS lookups via whois.com. Accepts URLs, bare domains, or whois.com lookup links, and returns registration dates, registrar, name servers, status codes, abuse contacts, registrant/technical contact details, and similar domain suggestions.
• 🧴 Incidecoder Scraper — Scrapes skincare product pages from INCIDecoder, extracting full ingredient lists, highlights, key/other ingredient groupings by function, per-ingredient ratings, and detailed explanations.
• 🎟️ Gametime Events Website Scraper — Scrapes event, venue, and ticket offer data from Gametime. Supports search, group/team, and single-event URLs, returning schedule, location, performer, and pricing details — including per-offer section, row, seat count, and deal rating — as structured JSON with pagination support.
• ⚖️ Austria Ediktsdatei Scraper — Scrapes Austrian judicial auction listings (Gerichtliche Versteigerungen) from edikte.justiz.gv.at, including property details, dates, valuations and PDF documents.

💬 Support & Contact

If you encounter any issues or have questions, please open an issue.

You can also find more of our actors on the Apify Store.

---

🇮🇩 Panduan Singkat (Bahasa Indonesia)

Actor ini mengambil artikel berita dari Detik.com secara otomatis.

Cara pakai:

1. Isi Start URLs dengan halaman indeks (misalnya https://finance.detik.com/indeks) atau langsung URL artikel. Jenis URL dideteksi otomatis.
2. Atur Max items per start URL untuk membatasi jumlah artikel per URL.
3. Aktifkan Scrape the whole site jika ingin menelusuri seluruh kanal utama Detik.
4. Isi Cache project name (opsional) agar artikel yang sudah diambil pada run sebelumnya dilewati.
5. Klik Start, lalu unduh hasilnya dalam format JSON, CSV, atau Excel.

Data yang diambil: judul, ringkasan, penulis, tanggal terbit dan perubahan, kanal/kategori, kata kunci, gambar, serta isi lengkap artikel.

Catatan: actor ini hanya mengambil data yang memang tampil untuk publik di Detik.com — tanpa login dan tanpa menembus konten berbayar.