thegauravgiri/nepali-news-dataset

🇳🇵 Nepali News Dataset & NLP Corpus

The comprehensive, open-access Nepali & English News Dataset for NLP and Machine Learning, automatically aggregated and updated every 4 hours.

• —Repository: thegauravgiri/nepali-news-dataset
• —Total Articles: 15,000+ full-text articles and growing
• —Update Frequency: Every 4 hours via automated GitHub Actions pipelines
• —Languages: Nepali (np / ne) and English (en) in clean UTF-8 Devanagari encoding
• —License: MIT License

⚡ Free Zero-Config API Endpoints

You can directly fetch real-time and historical news via Hugging Face raw endpoints with no authentication or rate limits:

1. Today's Live News (Updated Every 4 Hours)

GET https://huggingface.co/datasets/thegauravgiri/nepali-news-dataset/raw/main/data/today.json

2. Historical Daily Archive (YYYY-MM-DD.json)

GET https://huggingface.co/datasets/thegauravgiri/nepali-news-dataset/raw/main/data/2026-08-23.json

3. Serverless Dataset Query API

GET https://datasets-server.huggingface.co/rows?dataset=thegauravgiri%2Fnepali-news-dataset&config=default&split=train&offset=0&limit=100

💻 Usage with Hugging Face datasets Library

from datasets import load_dataset

# Load entire corpus out of the box
dataset = load_dataset("thegauravgiri/nepali-news-dataset")

# View the dataset structure
print(dataset)
print(dataset["train"][0])

Filter Nepali Articles

nepali_news = dataset["train"].filter(lambda row: row["language"] == "np")
print(f"Total Nepali articles: {len(nepali_news)}")

📊 Dataset Schema

Each row contains:

📰 Supported News Portals

📄 License

This dataset is published under the MIT License.