Oct 13, 2026
- [Using specialized data to improve research agent accuracy](https://parallel.ai/blog/specialized-data-labbench2)
Author: By Ethan Kam
October 14, 2026
# Web Search for Browserbase is now powered by Parallel
Parallel is now the default search provider for the Browserbase Search API and Browserbase Agents, so agents can navigate the web end to end, from discovery and research to action.
Tags:Product
Starting today, Parallel is the default search provider for the Browserbase Search API and Browserbase Agents, so agents can navigate the web end to end, from discovery and research to action.
## Why search comes before the browser
Most agent tasks start before the agent knows which page to open. It has to find sources, read the useful ones, and then open a browser for the steps that need one, such as logging in to a portal or working through a complex page. Browserbase’s docs describe this order as search, then fetch, then browsers. Search finds the starting points, fetch pulls content from most pages quickly, and a browser session handles the hard-to-reach data.
Search is the first step, so the later steps depend on its results. When search returns the right URLs, your agent opens fewer browser sessions and spends fewer tokens reading pages that don’t answer the question.
## What changes for Browserbase users
- - By default, requests to /v1/search and bb.search.web now use Parallel unless you choose another provider.
- - You can set provider on each request.
- - A mode parameter accepts advanced or fast[advanced or fast].
- - Parallel is also now the search provider behind Browserbase’s Agents products.
## Getting started
Use your existing Browserbase API key:
123456789curl https://api.browserbase.com/v1/search \
-H "Content-Type: application/json" \
-H "x-bb-api-key: $BROWSERBASE_API_KEY" \
-d '{
"query": "headless browser infrastructure",
"numResults": 5,
"mode": "advanced"
}'``` curl https://api.browserbase.com/v1/search \ -H "Content-Type: application/json" \ -H "x-bb-api-key: $BROWSERBASE_API_KEY" \ -d '{ "query": "headless browser infrastructure", "numResults": 5, "mode": "advanced" }' ``` 12345678910import Browserbase from "@browserbasehq/sdk";
const bb = new Browserbase();
const res = await bb.search.web({
query: "headless browser infrastructure",
numResults: 5,
mode: "fast" // or "advanced"
});``` import Browserbase from "@browserbasehq/sdk"; const bb = new Browserbase(); const res = await bb.search.web({ query: "headless browser infrastructure", numResults: 5, mode: "fast" // or "advanced"}); ``` Read Browserbase’s announcement[announcement] and the Search docs[Search docs] for more information.
To use Parallel Search directly, start at docs.parallel.ai[docs.parallel.ai].
### About Browserbase
Browserbase is the complete platform to build and deploy agents that browse the web like humans. One API key gives agents Browsers, Search, Fetch, and Agent Identity to act where APIs can't.
### About Parallel
Parallel Web Systems builds web search and research APIs for AI agents. Developers and enterprises use Parallel to give their agents accurate, fresh, and verifiable information from the open web, through one API suite that covers search, extraction, deep research, entity discovery, and continuous monitoring. Parallel runs its own web-scale index and returns token-efficient outputs designed for LLMs, with citations, reasoning, and confidence scores on every research result.
