ddgs-search

Multi-engine web search without API keys is possible by querying Google, Bing, DuckDuckGo, Brave, Yandex, Yahoo, Wikipedia, and Mojeek through a free ddgs-based backend. Results are returned as JSON containing titles, URLs, and snippets.

Finding arXiv papers for a literature review is done via the free arXiv API, which returns paper metadata including titles, abstracts, authors, categories, and dates directly in a structured JSON format for immediate consumption.

You can collect evidence and compare approaches across different search engines with ddgs-search. It applies a multi-engine discovery process to gather current sources and outputs structured JSON results for direct comparison.

The ddgs dependency supports producing web-search-plus-compatible JSON output. This allows other tools and skills to consume the titles, URLs, snippets, and paper metadata directly without additional formatting.

Limitations of a free multi-engine search backend include relying on the ddgs dependency for web queries and the arXiv API for papers, which may restrict result volume or rate limits compared to paid key-based search APIs.

To get current web sources and arXiv paper metadata in a single JSON output, use a multi-engine search skill that combines ddgs-based web discovery with arXiv API querying. It outputs titles, URLs, snippets, and academic metadata.