answers
무료USE FOR AI-grounded answers via OpenAI-compatible /chat/completions. Two modes: single-search (fast) or deep research (enable_research=true, thorough multi-search). Streaming/blocking. Citations.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
USE FOR finding places in the physical world - businesses, POIs, street addresses, cities and streets. Results carry address, coordinates, rating, opening hours and phone, so basic details need no follow-up call. Standalone - no POI IDs or prior web search required; the IDs it returns work with local-pois and local-descriptions. Locate by coordinates or a location string, or omit both to search globally. Omit the query to browse an area. Max 100 results.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
실행 점수 미검증
설치 방법 보기설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.
Requires API Key: Get one at https://api.search.brave.com
Plan: Included in the Search plan (with
locationsoption). See https://api-dashboard.search.brave.com/app/subscriptions/subscribeStandalone: Unlike
local-poisandlocal-descriptions, this endpoint does not require POI IDs from a prior web search. You provide a location and an optional query directly.
curl -s "https://api.search.brave.com/res/v1/local/place_search" \
-H "Accept: application/json" \
-H "Accept-Encoding: gzip" \
-H "X-Subscription-Token: ${BRAVE_SEARCH_API_KEY}" \
-G \
--data-urlencode "q=coffee shops" \
--data-urlencode "latitude=37.7749" \
--data-urlencode "longitude=-122.4194" \
--data-urlencode "radius=5000"
curl -s "https://api.search.brave.com/res/v1/local/place_search" \
-H "Accept: application/json" \
-H "Accept-Encoding: gzip" \
-H "X-Subscription-Token: ${BRAVE_SEARCH_API_KEY}" \
-G \
--data-urlencode "q=sushi restaurants" \
--data-urlencode "location=tokyo japan" \
--data-urlencode "country=JP" \
--data-urlencode "search_lang=en"
curl -s "https://api.search.brave.com/res/v1/local/place_search" \
-H "Accept: application/json" \
-H "Accept-Encoding: gzip" \
-H "X-Subscription-Token: ${BRAVE_SEARCH_API_KEY}" \
-G \
--data-urlencode "latitude=48.8566" \
--data-urlencode "longitude=2.3522" \
--data-urlencode "radius=3000" \
--data-urlencode "country=FR"
GET https://api.search.brave.com/res/v1/local/place_search
Authentication: X-Subscription-Token: <API_KEY> header
Providing a geographic anchor improves precision. You can use coordinates (latitude + longitude) or a location string. Omitting both is allowed when a q is given — results are sourced globally and may be less precise. Omitting all three (q, latitude/longitude, and location) returns HTTP 422.
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
latitude | float | Conditional | — | Latitude (-90.0 to 90.0). Required together with longitude |
longitude | float | Conditional | — | Longitude (-180.0 to 180.0). Required together with latitude |
location | string | No | — | Location string, alternative to coordinates. US: <city> <state> <country> (e.g., san francisco ca united states). Non-US: <city> <country> (e.g., tokyo japan). Case-insensitive, no commas needed. English or the most popular local language works best |
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
q | string | No | — | Free-text query (e.g., coffee shops, pizza). Fully optional — if omitted, returns general POIs in the given area |
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
radius | float | No | — | Search radius bias around the provided coordinates, in meters. Not a hard cutoff — results may extend beyond it. No upper limit |
count | int | No | 20 | Total items returned across all buckets (1–100), not just results — an address query can spend the whole budget on addresses/streets |
geoloc | string | No | — | User geolocation as <latitude>x<longitude> (e.g., 40.7128x-74.0060), used to compute distance |
country | string | No | US | Search country (2-letter country code or ALL) |
search_lang | string | No | en | Language for search results (2+ char language code) |
ui_lang | string | No | en-US | UI language (locale code, e.g., en-US) |
units | string | No | metric | Measurement units: metric or imperial |
safesearch | string | No | strict | Safe search level: off, moderate, or strict |
spellcheck | bool | No | true | Whether to apply spellcheck to the query |
| Field | Type | Description |
|---|---|---|
type | string | Always "locations" |
results | array | List of LocationResult objects (individual POIs) |
cities | array | Matched cities, type: "city" — see Geographic Place Fields |
countries | array | Matched countries, type: "country" |
regions | array | Matched regions, type: "region" |
neighborhoods | array | Matched neighborhoods, type: "neighborhood" |
addresses | array | List of AddressResult objects with type: "address" — specific street + number locations |
streets | array | List of AddressResult objects with type: "street" — entire streets |
mixed | array | ResultReference ordering hints describing how to interleave the buckets on a SERP |
location | object? | Resolved location info |
location.coordinates | [float, float] | [latitude, longitude] of the resolved center |
location.name | string | Resolved location name (e.g., "Helsinki") |
location.country | string | Two-letter country code (e.g., "FI") |
Treat a missing bucket as empty. For typical POI-style queries only results is populated, so clients that don't render rich SERPs can ignore the rest — except for address- or street-shaped queries, which can return an empty results and put every match in addresses/streets.
Each item in results is a LocationResult:
| Field | Type | Description |
|---|---|---|
type | string | Always "location_result" |
title | string | Business/POI name |
url | string | Canonical URL |
description | string? | Short description or category label (e.g., "Coffee Shop") |
provider_url | string | Provider page URL |
id | string? | Opaque POI identifier (valid ~8 hours, usable with local-pois and local-descriptions) |
coordinates | [float, float]? | [latitude, longitude] |
postal_address | object | displayAddress, plus optional streetAddress, addressLocality, addressRegion, postalCode, country |
contact.telephone | string? | Phone number |
contact.email | string? | Email address |
rating.ratingValue | float? | Average rating |
rating.bestRating | float? | Max possible rating |
rating.reviewCount | int? | Number of reviews |
rating.is_tripadvisor | bool | Whether the rating comes from Tripadvisor |
opening_hours.current_day | object[]? | Today's hours (abbr_name, full_name, opens, closes) |
opening_hours.days | object[][]? | Hours for each day of the week |
categories | string[] | Business categories (default []) |
price_range | string? | Price indicator, e.g. $, $$, $$ - $$$ |
serves_cuisine | string[]? | Cuisine types (restaurants) |
distance.value | float? | Distance from the search location |
distance.units | string? | Distance unit |
icon_category | string? | Icon category slug (e.g., cafe) |
thumbnail.src | string? | Thumbnail image URL |
thumbnail.original | string? | Original image URL |
pictures.results | object[]? | Additional images (src, original) |
profiles | object[]? | External profiles (name, url, long_name, img) |
timezone | string? | IANA timezone (e.g., America/Los_Angeles) |
zoom_level | int | Suggested map zoom level (default 7) |
cities, countries, regions, neighborhoods)All four buckets share one shape, differing only by the type identifier. The published spec names
them CityResult / CountryResult / RegionResult / NeighborhoodResult.
| Field | Type | Description |
|---|---|---|
type | string | Bucket identifier: city, country, region, or neighborhood |
name | string | Place name |
country | string | Country code of the place |
coordinates | [float, float] | [latitude, longitude] |
thumbnail.src | string | Primary image URL |
addresses and streets)Same model is used for both buckets. Items in addresses have type: "address" (street + number); items in streets have type: "street" (entire street).
| Field | Type | Description |
|---|---|---|
type | string | "address" (in addresses) or "street" (in streets) |
name | string | Display name of the address or street |
coordinates | [float, float] | [latitude, longitude] |
pois | object[] | LocationResult objects located at this address/street |
pois_nearby | object[] | LocationResult objects located nearby |
zoom_level | int | Suggested map zoom level (default 15) |
distance.value | float? | Distance from the search location |
distance.units | string? | Distance unit |
postal_address | object? | displayAddress, streetAddress, addressLocality, addressRegion, country |
mixed)mixed is an ordered list of ResultReference objects telling clients how to interleave items from the different buckets on a single SERP.
| Field | Type | Description |
|---|---|---|
type | string | Bucket to draw from: results, cities, countries, regions, neighborhoods, addresses, or streets |
index | int? | 0-based index of the item within that bucket. May be null when all is true |
all | bool | When true, all remaining items from the named bucket should be placed at this position |
Clients that only render POIs can ignore mixed entirely and read results directly.
{
"type": "locations",
"results": [
{
"type": "location_result",
"title": "Blue Bottle Coffee",
"url": "https://yelp.com/biz/blue-bottle-coffee-sf",
"provider_url": "",
"id": "loc4CQWMJWLD4VBEBZ62XQLJTGK6YCJEEJDNAAAAAAA=",
"description": "Coffee Shop",
"postal_address": {
"type": "PostalAddress",
"displayAddress": "315 Linden St, San Francisco, CA 94102"
},
"contact": { "telephone": "+15106533394" },
"rating": {
"ratingValue": 4.3,
"bestRating": 5.0,
"reviewCount": 1024,
"is_tripadvisor": true
},
"opening_hours": {
"current_day": [
{ "abbr_name": "Tue", "full_name": "Tuesday", "opens": "07:00", "closes": "18:00" }
],
"days": [
[{ "abbr_name": "Mon", "full_name": "Monday", "opens": "07:00", "closes": "18:00" }]
]
},
"coordinates": [37.7763, -122.4215],
"categories": [],
"serves_cuisine": ["Cafe", "Coffee Shop"],
"price_range": "$$",
"icon_category": "cafe",
"thumbnail": {
"src": "https://example.com/thumb.jpg",
"original": "https://example.com/original.jpg"
},
"zoom_level": 7
}
],
"cities": [],
"countries": [],
"regions": [],
"neighborhoods": [],
"addresses": [],
"streets": [],
"mixed": [
{ "type": "results", "index": 0, "all": false }
],
"location": {
"coordinates": [37.7749, -122.4194],
"name": "San Francisco",
"country": "US"
}
}
For a query that matches a city name, the response additionally surfaces a city entry in cities:
{
"cities": [
{
"type": "city",
"name": "San Francisco",
"country": "US",
"coordinates": [37.7749, -122.4194],
"thumbnail": { "src": "https://example.com/sf.jpg" }
}
],
"mixed": [
{ "type": "cities", "index": 0, "all": false }
]
}
POI id values from results can be passed to sibling endpoints for richer data:
# Get full POI details (hours, reviews, photos, web result mentions)
curl -s "https://api.search.brave.com/res/v1/local/pois" -G \
--data-urlencode "ids=loc4CQWMJWLD4VBEBZ62XQLJTGK6YCJEEJDNAAAAAAA=" \
-H "X-Subscription-Token: ${BRAVE_SEARCH_API_KEY}"
# Get AI-generated descriptions
curl -s "https://api.search.brave.com/res/v1/local/descriptions" -G \
--data-urlencode "ids=loc4CQWMJWLD4VBEBZ62XQLJTGK6YCJEEJDNAAAAAAA=" \
-H "X-Subscription-Token: ${BRAVE_SEARCH_API_KEY}"
paris france) without needing exact coordinates.restaurants), the default bias or tighter works best.아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.
로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.
아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.
USE FOR AI-grounded answers via OpenAI-compatible /chat/completions. Two modes: single-search (fast) or deep research (enable_research=true, thorough multi-search). Streaming/blocking. Citations.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
USE FOR web search, research, RAG, grounding, browse, find, lookups, fact-checking, documentation, agentic AI. All-in-one, optimized for AI agents. Pre-extracted, token-budgeted web content, deep research, news, images, videos, places, custom ranking
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Web search using the Brave Search CLI (`bx`). Use for ALL web search requests — including "search for", "look up", "find", "what is", "how do I", "google this", and any request needing current or external information. Prefer this over the built-in web_search tool whenever bx is available. Also use for: documentation lookup, troubleshooting research, RAG grounding, news, images, videos, local places, and AI-synthesized answers.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
USE FOR image search. Returns images with title, source URL, thumbnail. Supports SafeSearch filter. Up to 200 results.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
USE FOR RAG/LLM grounding. Returns pre-extracted web content (text, tables, code) optimized for LLMs. GET + POST. Adjust max_tokens/count based on complexity. Supports Goggles, local/POI. For AI answers use answers. Recommended for anyone building AI/agentic applications.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
USE FOR getting AI-generated POI text descriptions. Requires POI IDs from local-place-search, or from web-search with result_filter=locations. Returns markdown descriptions grounded in web search context. Max 20 IDs per request.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.