Brave updates its Search API with a new LLM context endpoint for AI and new pricing

Brave has announced a major overhaul of its Search API, introducing the LLM Context endpoint as its centerpiece, new developer-focused tools, and revised pricing. The LLM Context API returns web context in a compact format intended for LLM use rather than traditional browsing, so basically AI apps can pull relevant information from web pages instead of just a list of links. Brave says the endpoint is already powering more than 22 million answers per day and is used by its Ask Brave feature.

Brave describes a data first flow where queries run on Brave’s independent index and top pages are processed in real time. The system extracts and ranks smart chunks including text snippets, markdown, structured data such as JSON LD and tables, code context, forum discussions, and YouTube captions, then compiles the output per query. Developers can adjust parameters like token limits, output size, ranking behavior, and the number of URLs returned, with support for Goggles based filtering and localized results.

The release also adds API Skills and an API Assistant in the Developer Portal, and reorganizes access under public plans including Search, Answers, Spellcheck, and Autocomplete, with $5 in monthly free credit per plan when attribution is provided. Brave also highlights enterprise oriented options like SOC 2 Type II alignment and Zero Data Retention, and says it does not use customer queries to train LLMs.