Indian News Watcher

Pull recent news from named Indian publications across English, Hindi, Kannada and Tamil, and get back a clean list of articles — each with its source, language, date, link, summary, and a cross-language mood score (positive / negative / neutral).

Most news scrapers just wrap Google News in English. This one covers named Indian sites in multiple languages and — the part that sets it apart — scores the mood of each article in its own language, including Hindi and Kannada, where simple English sentiment tools fail.

Who uses it: teams and builders who need Indian news data at scale — brand and topic monitoring across languages, regional-sentiment dashboards, media research, and feeding pre-labelled multilingual news into an LLM or analytics pipeline.

What it does

• Reads each publication's RSS feed (stable, rarely breaks) — not fragile page scraping — so upkeep stays near zero.
• Filters by language, source, keyword, and date range.
• Removes duplicate stories (the same headline often runs on many sites).
• Optionally fetches the full article text.
• Optionally adds a mood score in any language, powered by Google Gemini.
• Returns one clean row per article.

Sources covered (v1)

Language	Publications
English	Times of India, The Hindu, NDTV, Hindustan Times, The Indian Express, Livemint
Hindi	NDTV India, Amar Ujala, Dainik Bhaskar
Kannada	Vijaya Karnataka, Prajavani, OneIndia Kannada
Tamil	OneIndia Tamil

Need another language or source? Telugu, Bengali, Marathi, Malayalam, or a specific publication — message me and I'll add it. This list grows based on what users actually ask for.

Input

Field	Meaning
languages	Which languages to include (default: English, Hindi, Kannada)
sources	Optional — limit to specific publications
keywords	Optional — keep only articles matching these words
dateFrom / dateTo	Optional — date range
maxItems	Cap on articles returned (also caps cost); default 200
includeBody	Fetch full article text (default off)
moodScore	Add a mood score (default on)
aiApiKey	Optional — run the mood score on your own Gemini key

Example

{
  "languages": ["Kannada", "Hindi"],
  "keywords": ["ಬೆಂಗಳೂರು", "मुंबई"],
  "maxItems": 50,
  "moodScore": true
}

Returns up to 50 recent Kannada/Hindi articles mentioning those terms, each like:

{
  "title": "ಬೆಂಗಳೂರಿನಲ್ಲಿ ಹೊಸ ಮೆಟ್ರೋ ಮಾರ್ಗ ಉದ್ಘಾಟನೆ",
  "link": "https://kannada.oneindia.com/...",
  "source": "OneIndia Kannada",
  "language": "Kannada",
  "category": "top",
  "publishedAt": "2026-07-01T09:15:00.000Z",
  "summary": "...",
  "mood": "positive",
  "moodReason": "New metro line opens, benefiting commuters"
}

The mood score, and its cost

The mood score is the reason to pick this over a generic news scraper. It runs on Google Gemini 2.5 Flash-Lite, which handles Indian languages well and costs a fraction of a cent per article (only the title and summary are sent). It is on by default and works out of the box — no setup. If a model call fails it retries and falls back automatically; if it still can't score an article, that article is returned with an empty mood rather than failing the run.

Heavy users can plug in their own Gemini key via aiApiKey to run mood scoring on their own account.

Pricing

Pay per article returned. Runs with the mood score enabled are priced slightly higher to cover the AI cost; article-only runs are cheaper.

Notes on fair use

This tool returns headlines, short summaries, links and dates — the same data publications put in their public RSS feeds. The optional full-text feature is for your own reading and analysis; don't republish full articles as your own.

Keywords

Indian news API, Hindi news scraper, Kannada news feed, regional Indian news monitor, India news sentiment, multilingual news API, Tamil news scraper, India news mood score.

---

Running / developing

• Mood key: set GEMINI_API_KEY as an actor secret (Apify Console → Actor → Environment variables) so the built-in scorer works for all users.
• Local mood test: GEMINI_API_KEY=... node src/mood.js runs a live English/Hindi/Kannada check; without the key it still checks the fail-soft path.
• Local feed test: node rss-test.mjs confirms every source feed still loads.
• Local full run: apify run (needs apify-cli). Charging events used: article (mood off) and article-with-mood (mood on) — define these in the Console pay-per-event pricing.