Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.browserbase.com/llms.txt

Use this file to discover all available pages before exploring further.

n8n is a fair-code licensed workflow platform that connects hundreds of services through a visual, node-based editor. The Browserbase node brings browser agents, Search, and Fetch into n8n so your workflows can gather context, inspect pages, and act on the web. Use the Agent resource to run Stagehand-powered browser agents in cloud-hosted Browserbase sessions. Use Search and Fetch to collect web context before launching a browser session, or to handle lightweight retrieval tasks on their own. Results flow into the rest of your n8n workflow.

​
Key features

No-code browser agents

Provide a URL and an instruction in plain English. Browser agents handle navigation, clicking, typing, and extraction from a visual workflow builder.

AI agent modes

Choose between CUA (vision-based), DOM (selector-based), or Hybrid mode depending on your use case and performance needs.

Search and Fetch resources

Find relevant URLs with Search or retrieve page content with Fetch without creating a browser session.

Built-in browser features

Use Verified browsers, residential proxies, CAPTCHA solving, ad blocking, session recording, and multi-region support.

Secure variables

Pass sensitive data like passwords and API keys with %variable% placeholders. The LLM never sees the values.

400+ integration ecosystem

Chain browser agent results with n8n’s extensive node library. Send extracted data to Slack, store it in Google Sheets, trigger webhooks, and more.

​
Agent modes

The Agent resource supports three modes that determine how the browser agent interacts with web pages:
ModeDescriptionBest for
CUAComputer Use Agent, uses vision and screen coordinatesComplex UIs, visual interactions, dynamic content
DOMUses DOM selectors, works with any LLMSpeed, simple pages, structured content
HybridCombines vision and DOM approachesFallback reliability, mixed content types

​
Model configuration

The Agent resource requires two model selections:
  • Driver Model: Powers the browser session (navigation, DOM interactions)
  • Agent Model: Orchestrates the high-level browser agent task
You can use Model Gateway with only your Browserbase API key instead of adding a separate model provider key. Model Gateway lets you switch between supported models from the node. Browserbase bills model usage at market rate through your Browserbase account.
If you bring your own Model API Key, both the Driver Model and Agent Model must be from the same provider. For example, if you select a Gemini driver model, your agent model must also be a Google model, and your Model API Key must be a Google API key.
Supported providers: Google (Gemini), OpenAI (GPT-4o, Computer Use Preview), and Anthropic (Claude).

​
What you can build

  • Research pipelines: Search the web, Fetch candidate pages, and send promising targets to a browser agent
  • Data extraction flows: Extract product prices, news headlines, or job listings, and pipe them into databases or spreadsheets
  • Form completion agents: Fill out and submit forms across multiple sites as part of a larger workflow
  • Competitive intelligence: Track competitor pricing, features, or content changes on a schedule
  • Lead enrichment: Visit company websites to gather contact info, tech stacks, or social profiles
  • Web flow validation: Validate web flows and capture results alongside your CI/CD pipeline

​
Get started

Quickstart guide

Install the community node, configure credentials, and run your first browser agent workflow.