Turkish Defense & Tech News Scraper — Savunma Sanayi + Milli Teknoloji

Scrape Turkish defense-industry and national-technology news with the parts other news scrapers leave out: full article text, the article's own images, publish date, author and automatic tagging of defense companies and weapon systems.

Built for OSINT analysts, defense-industry market research, journalists, and AI agents that need Turkish-language source material — not just headlines.

What you get per article

Field	Meaning
title	Headline (as published)
url	Publisher URL
source / sourceKey	Publisher name (e.g. SavunmaSanayiST)
publishedAt / publishedRaw	Publish date, ISO where available
author	Author or outlet account
summary	Lead / meta description
text	Full article text (Turkish)
imageUrl / images[] / imageCount	Lead image plus all in-article images
companies	Tagged companies/agencies: ASELSAN, BAYKAR, TUSAŞ, ROKETSAN, HAVELSAN, STM, OTOKAR, BMC, TÜBITAK SAGE, ASFAT, MKE, TEI, SSB, MSB, TOGG …
systems	Tagged systems/programmes: KAAN, AKINCI, BAYRAKTAR TB2/TB3, KIZILELMA, HÜRJET, HÜRKUŞ, GÖKBEY, ANKA, AKSUNGUR, MİLDEN, ALTAY, SİPER, HİSAR, GÖKDENİZ, ATMACA, ÇELİK KUBBE, İMECE …
topics	Theme tags: ihracat, seri-uretim, insansiz-hava, hava-savunma, uzay, test-basarisi, sozlesme, milli-teknoloji …
language, wordCount, scrapedAt	Basics for pipelines

Tagging is language-aware (Turkish casing is handled: İ, I, ı, Ş, Ğ …), so ASELSAN matches ASELSAN, ASELSAN'ın and ASELSAN’a alike.

Sources

Verified with live HTTP checks — status, feed availability and image extraction:

Key	Publisher	Type	Topic
savunmasanayist	SavunmaSanayiST	RSS	defense
millisavunma	Milli Savunma	RSS	defense
aasavunma	Anadolu Ajansı (Savunma)	RSS + URL filter	defense
ssb	Savunma Sanayii Başkanlığı (official)	HTML	defense
kokpit	Kokpit Aero	HTML	defense/aviation
gdh	GDH Digital	HTML	defense
veryansin	Veryansın TV	HTML	defense
defenceturk	DefenceTurk	HTML (proxy)	defense
haberturk_tek	Habertürk Teknoloji	HTML	milli teknoloji
tubitak	TÜBİTAK	HTML	milli teknoloji
teknofest	TEKNOFEST	HTML	milli teknoloji

Some publishers block datacenter IPs (DefenceTurk answers 403 without a proxy). That is exactly why this runs on Apify: switch on Apify Proxy and those sources work too. On overseas datacenters the ssb, tubitak and teknofest sources may additionally be geo-restricted — use a Turkish proxy group if a source returns nothing.

Input example

{
  "preset": "savunma",
  "keywords": ["KAAN", "İHA", "ihracat"],
  "maxItems": 50,
  "includeFullText": true,
  "includeImages": true,
  "dateFrom": "2026-09-01",
  "proxyConfiguration": { "useApifyProxy": true }
}

Leave keywords empty to get everything, or set preset to teknoloji for national technology, or alle for both lanes.

Output example

{
  "title": "DELTAV Ramjet Motorlu Füze",
  "url": "https://www.millisavunma.com/deltav-ramjet-motorlu-fuze/",
  "source": "Milli Savunma",
  "publishedAt": "2026-06-10T22:50:17+00:00",
  "text": "Türkiye'nin savunma ve uzay teknolojileri alanındaki özel girişimlerinden biri olan DeltaV …",
  "imageUrl": "https://www.millisavunma.com/wp-content/uploads/2026/06/deltav_ramjet_01.jpg",
  "images": ["…/deltav_ramjet_01.jpg", "…/deltav_ramjet_03.jpg", "…/deltav_ramjet_05.jpg"],
  "imageCount": 5,
  "companies": ["ASELSAN"],
  "systems": ["ATMACA"],
  "topics": ["hava-savunma", "test-basarisi"],
  "wordCount": 668,
  "language": "tr"
}

Use cases

• OSINT & defense monitoring — daily picture of the Turkish defense industry (procurement, exports, tests, programme milestones) with the source text for citations.
• Market & supplier research — who announces what, exports, capacity expansions, contracts.
• AI agents & RAG — Turkish-language corpus with images and entity tags, ready for embedding.
• Media & content pipelines — headline, lead photo and full text per story.

Pricing

Pay per event: you are charged for each article delivered to the dataset. Nothing is charged for articles that are filtered out (too short, keyword mismatch, duplicate).

Notes & limits

• Text and images are extracted from the publishers' own public pages; copyright stays with the publisher. Check your own use case (quoting, monitoring, research) against their terms.
• Paywalled or login-only articles are not bypassed.
• The scraper is server-side (no browser): fast and cheap. Sources that need JavaScript or that hard-block datacenter IPs are best served with Apify Proxy enabled.
• New publishers get added; the source registry is data, not code, so extensions are cheap.

Development

pip install -r requirements.txt
python3 -m src        # local run, input from storage/key_value_store/INPUT.json
apify run             # official local runner
apify push            # build & deploy to Apify