Search the web, images, videos, news, and local places through Brave's independent index — plus AI summarization and RAG-ready context extraction.
A Model Context Protocol (MCP) server that exposes the Brave Search API's search endpoints for web, local, video, image, news, and place search, along with AI summarization and LLM-context extraction.
Overview
The Brave Search MCP Server provides:
Comprehensive web search with rich metadata — results can include FAQ, Discussions, News, and Video items alongside standard web results
Local business and points-of-interest (POI) search with ratings, addresses, phone numbers, hours, and AI descriptions, with a graceful fallback to web results when enriched local data isn't available
Dedicated video, image, and news search endpoints
A structured place-search endpoint for POI data anchored to a location or lat/long coordinates
AI-generated summarization of web search results via Brave's Summarizer API
Pre-extracted, relevance-ranked web content via Brave's LLM Context API — actual page substance (text chunks, tables, code blocks) rather than just links and snippets
Perfect for:
AI agents that need general-purpose web search grounded in an independent search index
Local business lookup and "near me" style assistants
News monitoring and current-events tools
Media search integrations (images, videos)
RAG pipelines that need pre-extracted, relevance-ranked web content for grounding
Quickly summarizing a set of search results into an AI-generated answer
Tools
Performs web searches using the Brave Search API and returns comprehensive search results with rich metadata.
When to use:
General web searches for information, facts, or current topics
Location-based queries (restaurants, businesses, points of interest)
News searches for recent events or breaking stories
Finding videos, discussions, or FAQ content
Returns a JSON list of web results with title, description, and URL. When result_filter is empty, results may also contain FAQ, Discussions, News, and Video items.
Searches for images via the Brave Search API. Returns direct image URLs, source pages, and dimensions. Images are returned as URLs — no base64 encoding.
Retrieves points of interest (POIs) with structured business data via Brave's dedicated place-search endpoint. Returns addresses, hours, ratings, categories, and contact info.
Geographic context is required — provide latitude/longitude or a location string (e.g. 'san francisco ca united states').
Access requires a Brave Search API Pro plan.
Inputs:
- `query` (string, required) — Search query — shapes result type, e.g. 'coffee shops' or 'Eiffel Tower'- `location` (string, optional) — Location context, e.g. 'san francisco ca united states' or 'tokyo japan'- `latitude` (float, optional) — Latitude (-90 to 90)- `longitude` (float, optional) — Longitude (-180 to 180)- `radius` (int, optional) — Proximity bias in metres (not a hard cutoff)- `count` (int, optional, default: 20) — Results to return (1–50)- `country` (CountryCode, optional, default:
Retrieves an AI-generated summary of web search results using Brave's Summarizer API.
Workflow: call search_web with summary=true first, then pass the returned summarizer_key to this tool.
Requires a Brave Search API Pro AI subscription.
Inputs:
- `key` (string, required) — Summarizer key from search_web called with summary=true- `entity_info` (bool, optional, default: false) — Include related entity information
output data schema:
{ text: string;}
Retrieves pre-extracted, relevance-ranked web content using Brave's LLM Context API, optimised for AI agents, LLM grounding, and RAG pipelines.
Unlike a web search (links + short descriptions), this tool returns the actual substance of matching pages — text chunks, tables, code blocks — so the model can reason over it directly.
When to use:
Grounding answers in fresh, relevant web content (RAG)
Question answering and fact-checking against current sources
Gathering source material without manually fetching pages
When relaying results in markdown environments, cite source URLs from the 'sources' map.
Inputs:
- `query` (string, required) — Search query (max 400 chars, 50 words)- `country` (CountryCode, optional, default: "US") — Country for results- `search_lang` (SearchLang, optional, default: "en") — Search language- `count` (int, optional, default: 20) — Number of results to consider (1–50
API Parameters Reference
Every tool returns the same top-level envelope. Only data varies per tool.
- `x_loc_lat` (float, optional) — User latitude (-90 to 90)
- `x_loc_long` (float, optional) — User longitude (-180 to 180)
- `x_loc_city` (string, optional) — User city
- `x_loc_country` (string, optional) — User 2-letter country code
output data schema:
{ // This model declares no fixed fields of its own — the entire response // shape is the passthrough JSON from Brave's LLM Context API. The tool // description references a `sources` map for citing result URLs. [key: string]: unknown;}
All data models accept additional, undocumented fields beyond what's listed above (extra="allow") — Brave may add fields to its API responses that aren't reflected in these schemas yet.
freshness
pd
pw
pm
py
YYYY-MM-DDtoYYYY-MM-DD
units (Units: metric | imperial) — Measurement units
CountryCode, SearchLang, and UiLang are each closed lists of ISO-style codes matching Brave's supported values — pass one of the enumerated codes for the given parameter.