webfetch.io

Webfetch Search API adds Brave Search results as JSON

Daniel Jones

The Webfetch Search API returns Brave Search results as JSON. Brave went live on October 20, 2026 as the fourth engine, after Google, Bing and Yahoo launched on October 16. One GET request returns a ranked page of Brave results.

GET https://api.webfetch.io/v1/brave

The endpoint takes the same parameters and returns the same response shape as the other three engines, so code written for one engine reads Brave by changing the path. Brave runs its own independent search index. ScrapingBee has no Brave endpoint.

What a Brave results page returns

  • Organic results. About 20 a page, each with its position, title, destination url, displayed_url, description, domain and the date Brave shows beside the snippet, when there is one.
  • The sponsored result. At most one, at the top, in top_ads, with its landing page in url and Brave's click URL in tracking_url.
  • Related searches. In related_searches.
  • People also ask, with the answers. Each questions entry carries the question in text and Brave's answer in answer. The Google, Bing and Yahoo endpoints return the questions only.

The local, map, image, news, knowledge graph and AI overview keys are present for compatibility and always empty.

Request

curl "https://api.webfetch.io/v1/brave?search=best+running+shoes&country_code=us" \
  -H "Authorization: Bearer $WEBFETCH_KEY"

The key goes in the Authorization: Bearer header, or in the api_key query parameter. The other parameters are the ones the Google, Bing and Yahoo endpoints take: page (1 to 50), light_request, search_type (classic for the full page or ads for the ads only), add_html to return the page HTML, and tag for your own label.

Brave chooses results by the location of the exit the search runs from, set with country_code. Brave has no language setting. A country with no search exit is refused with 400 country_unavailable, and the Brave Search API page lists the countries available now.

Response

A real Brave search for "best running shoes" from the US, captured October 20, 2026 and trimmed to the first organic result:

{
  "meta_data": {
    "url": "https://search.brave.com/search?q=best%20running%20shoes&source=web",
    "number_of_results": null,
    "location": null,
    "number_of_organic_results": 20,
    "number_of_ads": 1,
    "number_of_page": 1,
    "zero_results_for_original_query": false
  },
  "organic_results": [
    {
      "position": 1,
      "title": "7 Best Running Shoes in 2026 | RunRepeat",
      "url": "https://runrepeat.com/guides/best-running-shoes",
      "displayed_url": "runrepeat.com \u203a running shoes \u203a best running shoes",
      "description": "The 7 best running shoes based on extensive wear and lab testing. We award the best overall, best daily trainer, best for tempo, best for race, best shock absorption, best energy return and best stability.",
      "domain": "runrepeat.com",
      "date": "September 13, 2026",
      "date_utc": null,
      "rich_snippet": [],
      "sitelinks": []
    }
  ],
  ...
}

Brave showed no People also ask block for this query, so its questions list is empty. Every key is present on every response, and a list with nothing in it is an empty array, so a client never checks whether a key exists.

Credits

  • 10 credits for a light request, the default, and 15 with light_request=false. Every Brave search loads in a browser, and a light request skips images, fonts and media.
  • An ads-only search costs the same as a full results page.
  • A failed search costs nothing: a block, a timeout or a page that is not a results page is never billed.
  • A results page with no results is billed, because it is the engine's answer.

Credits come from the plan's monthly allowance or the free credit balance, at the same cost as Google, Bing and Yahoo. Plans are on the pricing page.

Access

Brave's robots.txt disallows automated access to its result pages, and webfetch honors robots.txt until the robots.txt opt-out is on for the key. Turning it on records your acceptance of the acceptable use policy, and every search is logged.

  • Paid plans turn the opt-out on in the key portal, and search works right away.
  • Free accounts request search access in the key portal. A person reviews and approves each request, usually within a day. Approved free accounts search one at a time from the free credit balance. Create a free account to request it.

Search tool, SDKs and MCP

The Brave search tool runs the same search in the browser. Pick the country, the results page, full results or ads only, and a light or full browser load, then download the results as JSON or CSV. A search there costs the same credits as the API.

The Webfetch Brave search tool with a saved search for best running shoes, showing the sponsored result at the top and the first organic results
The Brave search tool with its saved example search.

SDK 1.2.0 for JavaScript, Python and PHP adds brave(), which returns the parsed response:

// npm install @tuxxin/webfetch
const r = await client.brave('best running shoes');

# pip install webfetch-io
r = client.brave("best running shoes")

// composer require tuxxin/webfetch
$r = $client->brave('best running shoes');

The search_web tool on the MCP server at https://webfetch.io/mcp accepts engine "brave", so an AI agent can run Brave searches on the same key.

The Search API overview covers all four engines, and the full reference lists every parameter, header and error.

All posts