Open-source privacy-focused browser Brave today introduced a new "Summarizer" option for its dedicated search engine, Brave Search.
Powered by large language models (LLMs), the Summarizer is designed to offer up concise answers at the top of the Brave Search results page for any query. This is not a generative AI model like OpenAI, as the LLMs trained by Brave are processing multiple sources of information solely from web results.
Google Search has long offered a similar summarization feature, but it is typically based on a single result from a website chosen as an authoritative source rather than multiple websites.
Brave says that using web search results for the Summarizer provides an accurate answer that is expressed in coherent language. Data is also cited via links for accurate attribution of information and to allow users to assess the trustworthiness of sources. Responses are only as accurate as the websites that the browser uses for the summary, but each source is clearly defined.
The Summarizer will also provide improved search result descriptions by using a summary rather than a snippet of text pulled from a website.
Brave's Summarizer feature is available for all Brave Search users starting today, on both desktop and mobile. Brave users who don't want the Summarizer can turn it off in the settings.
More information on the Summarizer and how it works can be found on Brave's website.