Brave Search API

What it does

Brave Search API's LLM Context endpoint takes a search query and returns ranked, extracted passages from web pages, including text, tables, code, JSON, forum threads and video captions. You set a token budget, typically from about 1,000 to 32,768 tokens, to control how much content comes back. Unlike standard web search, which returns titles and snippets for people to read, this response is formatted for language models and agents.

Developers call the endpoint over HTTP with an API key in the X-Subscription-Token header, using GET or POST. Options include freshness filters, relevance threshold modes, safesearch, location headers for local results, and Goggles for re-ranking or limiting which domains ground the model. Each source can include its URL and timestamp, plus optional metadata such as site name and favicon.

This suits developers building AI agents, RAG pipelines, chatbots and fact-checking tools that need current web grounding. Access requires a registered API key from the Brave API dashboard. Larger token budgets return more context but raise model inference costs and latency.

What it can do

• Returns pre-extracted page text, tables, code and structured data for LLM consumption.
• Ranks relevant passages in a single call with configurable token and URL budgets.
• Filters by relevance threshold (strict, balanced, lenient, disabled) and freshness windows.
• Supports Goggles to re-rank results and define custom domain sets.
• Supports location-aware local results via location headers.
• Provides source metadata including URL and fetched timestamp, with optional site name, favicon and thumbnail.

Example workflows

Grounding a chatbot answer

Input: a user question and a token budget. Steps: send a GET request to the LLM Context endpoint with the question and budget; pass the returned passages and source URLs to the model. Output: an answer grounded in the cited pages.

Filtering recent sources for research

Input: a topic query with freshness set to pw. Steps: request LLM Context with the freshness filter and a strict threshold; check each source's timestamp. Output: a set of recent, traceable passages for review.

Before you choose

• Larger token budgets return more context but increase model inference costs and latency.
• Confirm current plan terms and rates on the pricing page before committing to a budget.

Listed under

• Coding
• MCP and agent tools
• AI agents
• Developer tools
• Grounding
• LLM grounding
• RAG
• Search API

Get started

Register in the Brave API dashboard, then send requests with your key in the X-Subscription-Token header.

1. Create an account at the Brave API dashboard and generate an API key.
2. Send a test request to https://api.search.brave.com/res/v1/llm/context with your key in the X-Subscription-Token header.
3. Set maximum_number_of_tokens and context_threshold_mode to match your task, then add freshness or Goggles if needed.

A $5 monthly free credit is stated, which is a time-limited allowance and not a permanent free tier. Two rates appear: $5 per 1,000 requests and $4 per 1,000 requests plus $5 per million input/output tokens. Check the pricing documentation for which applies to LLM Context.

Source: brave.com/search/api, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Brave Search API | Brave ↗brave.com
• Brave Search - API ↗api-dashboard.search.brave.com

Company website

brave.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.