Metaphor

Metaphor is a website built around a simple pitch: helping users rediscover the internet. Rather than returning the same handful of ranked pages, it is described as a search surface that reaches across different kinds of content, including Twitter, Wikipedia, news, and papers. That spread is the appeal, because a question that needs discussion, a reference, and a recent article can often be answered inside one search instead of three. Two features support refinement directly: website filtering, which lets you exclude sources you do not want, and autocomplete suggestions, which help you shape a query as you type. The listing also puts weight on the API, describing it as powerful enough that a developer can add web search to an application with only a few lines of code.

- Search across Twitter, Wikipedia, news, and papers at once.

- Use website filtering to cut sources you do not want.

- Rely on autocomplete suggestions to tighten a query.

- Add search to your own product through the API.

- Researchers gathering material across several content types.

- Developers who need a search feature inside an application.

- Readers who want discussion and context, not just official pages.

- Anyone frustrated by repetitive results on a general engine.

Any search service is only as good as its index, and coverage of a given topic can be thin, especially for niche or non English material. If you are building on the API, expect rate limits, usage based costs, and the possibility that endpoints change, so design your application to degrade gracefully when search is unavailable. The description says users can try it for free, but free access is an entry point rather than an unlimited plan, so read the current pricing before you depend on it for a product. Third party search platforms can also be acquired, repositioned, or shut down, which means anything you build on one should have a fallback. Finally, verify important claims against primary sources rather than trusting a results page.