News API Built for LLM Applications

Give your language models a steady stream of structured, real-world news data. Clean JSON, rich metadata, and 26M+ articles for training, fine-tuning, and live retrieval.

Why Currents works well with LLMs

Structured JSON

Every article includes title, description, full text, published date, source, language, and category — ready to drop into prompts or vector databases.

26M+ Articles

A large, diverse corpus for training, evaluation, or few-shot examples. Covering 70+ countries and 20+ languages over the past 4 years.

Real-Time Feed

90K+ new articles daily. Keep your retrieval pipeline fresh with the latest news without managing scrapers or parsers.

Common LLM use cases

RAG Pipelines

Chunk articles, embed them, and retrieve relevant context for question-answering. Currents gives you the raw text and metadata to build high-quality retrieval.

News Summarization

Feed article text into summarization models to generate briefs, alerts, or daily digests across topics and regions.

Sentiment & Topic Modeling

Use historical and real-time news to train classifiers, detect trends, and monitor public sentiment on companies, products, or events.

Agent Tooling

Give autonomous agents a search_news tool that queries live data. JSON responses are easy to parse into structured action outputs.

Power your LLM with live news data

Start free with 1,000 daily requests. No credit card required.