Train and evaluate browser agents with BrowserEnv | Browserbase | Browserbase

Skip to content

- Primitives
- Open Source
  - Browse CLIGive your agent browsing skills with a single command
  - StagehandThe most popular AI browser automation framework
- Resources
Pricing
Docs
Log in

Skip to content

- Primitives
- Open Source
  - Browse CLIGive your agent browsing skills with a single command
  - StagehandThe most popular AI browser automation framework
- Resources
Pricing
Docs
Log in

Browserbase

Primitives

Browsers
Web APIs
Runtime
Identity
Model Gateway
Observability
Stagehand
MCP

Industries

AI
Healthcare
Supply Chain
GTM
Tax
Legal

Developers

Docs
Templates
APIs & SDKs
Changelog
Status
Github

Company

Careers
Customers
Partner with Us
Trust & Security

Community

Twitter Linkedin Youtube

Privacy Policy
Terms of Service

Community

Twitter Linkedin Youtube Instagram

Privacy Policy
Terms of Service

Train and evaluate browser agents

Train browser agents on real websites with Prime Intellect’s RL pipelines, powered by Browserbase.

Read the docs Explore environments

Open the hosted training run

Play this video on YouTube

How it fits together

Cloud browsers

Browserbase

Cloud browser infrastructure that creates and manages browser sessions via API.

Environment layer

BrowserEnv

Environment abstraction that wraps Browserbase sessions for evaluation and training.

Prime orchestration

Prime CLI and hosted training

Orchestrates evaluation runs and RL training loops at scale.

Two modes, one environment

DOM mode

Structured DOM content and high-level Stagehand instructions.

Best for: instruction-following tasks, form filling, and navigation
Tools: navigate, observe, act, extract
Command: prime eval run browser-dom-example

Run DOM example

CUA mode

Screenshots and coordinate-based

Best for: visual, pixel-grounded tasks and complex UIs
Tools: click, type_text, scroll, screenshot
Command: prime eval run browser-cua-example

Run CUA example

Browserbase

Primitives

Browsers
Web APIs
Runtime
Identity
Model Gateway
Observability
Stagehand
MCP

Industries

AI
Healthcare
Supply Chain
GTM
Tax
Legal

Developers

Docs
Templates
APIs & SDKs
Changelog
Status
Github

Company

Careers
Customers
Partner with Us
Trust & Security

Community

Twitter Linkedin Youtube

Privacy Policy
Terms of Service

Community

Twitter Linkedin Youtube Instagram

Privacy Policy
Terms of Service

Train and evaluate browser agents

Train browser agents on real websites with Prime Intellect’s RL pipelines, powered by Browserbase.

Read the docs Explore environments

Open the hosted training run

Play this video on YouTube

How it fits together

Cloud browsers

Browserbase

Cloud browser infrastructure that creates and manages browser sessions via API.

Environment layer

BrowserEnv

Environment abstraction that wraps Browserbase sessions for evaluation and training.

Prime orchestration

Prime CLI and hosted training

Orchestrates evaluation runs and RL training loops at scale.

Two modes, one environment

DOM mode

Structured DOM content and high-level Stagehand instructions.

Best for: instruction-following tasks, form filling, and navigation
Tools: navigate, observe, act, extract
Command: prime eval run browser-dom-example

Run DOM example

CUA mode

Screenshots and coordinate-based

Best for: visual, pixel-grounded tasks and complex UIs
Tools: click, type_text, scroll, screenshot
Command: prime eval run browser-cua-example

Run CUA example

Instagram

Instagram