Brave Search API
An independent web search API for AI agents and RAG, with an LLM Context endpoint that returns ranked, compact passages in a single call.
About Brave Search API
The Brave Search API gives developers programmatic access to Brave's own web index, the same one behind Brave Search. Unlike many search APIs, it is not a wrapper around Google or Bing: Brave crawls, ranks and serves results from its own index, so long-tail queries are answered without a fallback to another engine. The company says the index covers 40+ billion pages and is used for search products, AI search, model training and agentic search.
For AI builders the key piece is the LLM Context endpoint. Where a normal results page returns titles and snippets for humans, LLM Context returns the best passages from the best pages, ranked and pre-formatted for a model, covering text, tables, code, JSON, forum threads and video captions. You can set a token budget from roughly 1k to 33k tokens and a relevance threshold (strict, balanced or lenient). Other endpoints cover classic Web, News, Image and Video search and an Answers endpoint that returns summarised, cited answers and is OpenAI SDK compatible. Goggles let you re-rank results or restrict them to a custom set of domains.
Brave positions the API for teams that want an independent index and predictable, prepaid billing: no subscription, free monthly credits, SOC 2 Type II attestation and Zero Data Retention options. Several of its performance claims (agent quality, latency, cost per correct answer) come from the AIMultiple test and Brave's own published runs, so treat them as vendor-reported and benchmark on your own queries. Storing results, for example to tune a model, requires a plan that explicitly grants storage rights.
Key features
Frequently asked questions
What is the Brave Search API?
It is a developer API for building apps on web data from Brave's independent search index, the same index behind Brave Search. It is used for search products, AI search, model training and agentic search.
How much does the Brave Search API cost?
The Search plan is listed at $5 per 1,000 requests with up to 50 queries per second. The Answers plan is $4 per 1,000 requests plus $5 per million input and output tokens, up to 2 queries per second. Billing is prepaid credits with no subscription, and Enterprise pricing is custom.
Is there a free tier for the Brave Search API?
Yes. Brave gives free monthly credits (the pricing page shows $5 on each of the Search and Answers plans, and the page header says up to $20 in free calls per month). A credit card is required as an anti-fraud measure but is not charged if you stay within the free credits.
What is the LLM Context endpoint?
LLM Context is Brave's agent-oriented endpoint. It returns the most relevant passages from the best pages, ranked and pre-formatted for a model, with a configurable token budget and relevance threshold, instead of a page of titles and snippets.
Is the Brave Search API a Google or Bing scraper?
No. Brave states that it is not a scraper repackaging Google or Bing results. It uses its own crawler, ranking models and index, which is also fed by opt-in contributions through the Web Discovery Project.
Can I store results from the Brave Search API?
Only with a plan that explicitly grants storage rights, for example if you want to train or tune an LLM on the results. Standard terms do not cover storing results.
How does Brave Search API compare with Tavily, Exa and Serper?
Brave's pitch is an independent index and low latency, and it cites AIMultiple tests ranking it first for agent quality. Those are vendor-promoted results, so test on your own queries. See our Tavily, Exa and Serper listings for the alternatives.
Reader endorsements 0 written
+ Write oneNo endorsements written yet. Be first.
Related in AI & ML 3 picks
All topicsCompare Brave Search API with…
Head-to-head with the competitors readers named.