Skip to content

Open-source Python library for AI browser agents, plus Browser Use Cloud: stealth browsers and a hosted agent API.

Feature examples

Magnus Müller55b433d74036
diff --git a/examples/features/custom_user_agent.py b/examples/features/custom_user_agent.py
deleted file mode 100644
index 9182efc29..000000000
--- a/examples/features/custom_user_agent.py
+++ /dev/null
@@ -1,73 +0,0 @@
-import argparse
-import asyncio
-import os
-import sys
-
-sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-from browser_use import Agent
-from browser_use.browser import BrowserProfile, BrowserSession
-from browser_use.controller.service import Controller
-from browser_use.llm import ChatAnthropic, ChatOpenAI
-
-
-def get_llm(provider: str):
- if provider == 'anthropic':
- return ChatAnthropic(model='claude-3-5-sonnet-20240620', temperature=0.0)
- elif provider == 'openai':
- return ChatOpenAI(model='gpt-4.1', temperature=0.0)
-
- else:
- raise ValueError(f'Unsupported provider: {provider}')
-
-
-# NOTE: This example is to find your current user agent string to use it in the browser_context
-task = 'go to https://whatismyuseragent.com and find the current user agent string '
-
-
-controller = Controller()
-
-
-parser = argparse.ArgumentParser()
-parser.add_argument('--query', type=str, help='The query to process', default=task)
-parser.add_argument(
- '--provider',
- type=str,
- choices=['openai', 'anthropic'],
- default='openai',
- help='The model provider to use (default: openai)',
-)
-
-args = parser.parse_args()
-
-llm = get_llm(args.provider)
-
-browser_session = BrowserSession(
- browser_profile=BrowserProfile(
- user_agent='foobarfoo',
- user_data_dir='~/.config/browseruse/profiles/default',
- )
-)
-
-agent = Agent(
- task=args.query,
- llm=llm,
- controller=controller,
- browser_session=browser_session,
- use_vision=True,
- max_actions_per_step=1,
-)
-
-
-async def main():
- await agent.run(max_steps=25)
-
- input('Press Enter to close the browser...')
- await browser_session.kill()
-
-
-asyncio.run(main())
diff --git a/examples/features/download_file.py b/examples/features/download_file.py
index 49ddfdc1c..2ac115135 100644
--- a/examples/features/download_file.py
+++ b/examples/features/download_file.py
@@ -9,33 +9,22 @@ from dotenv import load_dotenv
load_dotenv()
-from browser_use import Agent
-from browser_use.browser import BrowserSession
-from browser_use.llm import ChatGoogle
+from browser_use import Agent, BrowserProfile, BrowserSession, ChatGoogle
api_key = os.getenv('GOOGLE_API_KEY')
if not api_key:
raise ValueError('GOOGLE_API_KEY is not set')
-assert api_key is not None, 'GOOGLE_API_KEY must be set'
-llm = ChatGoogle(model='gemini-2.0-flash-exp', api_key=api_key)
+llm = ChatGoogle(model='gemini-2.5-flash', api_key=api_key)
-from browser_use.browser import BrowserProfile
-browser_session = BrowserSession(
- browser_profile=BrowserProfile(
- downloads_path='~/Downloads',
- user_data_dir='~/.config/browseruse/profiles/default',
- )
-)
+browser_session = BrowserSession(browser_profile=BrowserProfile(downloads_path='~/Downloads'))
async def run_download():
agent = Agent(
task='Go to "https://file-examples.com/" and download the smallest doc file.',
llm=llm,
- max_actions_per_step=8,
- use_vision=True,
browser_session=browser_session,
)
await agent.run(max_steps=25)
diff --git a/examples/features/drag_drop.py b/examples/features/drag_drop.py
deleted file mode 100644
index c5f920cdb..000000000
--- a/examples/features/drag_drop.py
+++ /dev/null
@@ -1,51 +0,0 @@
-import asyncio
-import os
-import sys
-
-sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-
-from browser_use import Agent
-from browser_use.llm import ChatGoogle
-
-api_key = os.getenv('GOOGLE_API_KEY')
-if not api_key:
- raise ValueError('GOOGLE_API_KEY is not set')
-
-# API key is automatically set from the environment variable GOOGLE_API_KEY
-llm = ChatGoogle(model='gemini-2.0-flash-exp')
-
-
-task_1 = """
-Navigate to: https://sortablejs.github.io/Sortable/.
-Then scroll down to the first examplw with title "Simple list example".
-Drag the element with name "item 1" to below the element with name "item 3".
-"""
-
-
-task_2 = """
-Navigate to: https://excalidraw.com/.
-Click on the pencil icon (with index 40).
-Then draw a triangle in the canvas.
-Draw the triangle starting from coordinate (400,400).
-You can use the drag and drop action to draw the triangle.
-"""
-
-
-async def run_search():
- agent = Agent(
- task=task_1,
- llm=llm,
- max_actions_per_step=1,
- use_vision=True,
- )
-
- await agent.run(max_steps=25)
-
-
-if __name__ == '__main__':
- asyncio.run(run_search())
diff --git a/examples/features/initial_actions.py b/examples/features/initial_actions.py
index 6ba530aa1..7e8d585c8 100644
--- a/examples/features/initial_actions.py
+++ b/examples/features/initial_actions.py
@@ -15,7 +15,6 @@ llm = ChatOpenAI(model='gpt-4.1-mini')
initial_actions = [
{'go_to_url': {'url': 'https://www.google.com', 'new_tab': True}},
{'go_to_url': {'url': 'https://en.wikipedia.org/wiki/Randomness', 'new_tab': True}},
- {'scroll_down': {'amount': 1000}},
]
agent = Agent(
task='What theories are displayed on the page?',
diff --git a/examples/features/multi-tab_handling.py b/examples/features/multi_tab.py
similarity index 83%
rename from examples/features/multi-tab_handling.py
rename to examples/features/multi_tab.py
index 7c8a8bb31..0f80f94f9 100644
--- a/examples/features/multi-tab_handling.py
+++ b/examples/features/multi_tab.py
@@ -19,7 +19,7 @@ from browser_use import Agent, ChatOpenAI
# video: https://preview.screen.studio/share/clenCmS6
llm = ChatOpenAI(model='gpt-4.1-mini')
agent = Agent(
- task='open 3 tabs with elon musk, trump, and steve jobs, then go back to the first and stop',
+ task='open 3 tabs with elon musk, sam altman, and steve jobs, then go back to the first and stop',
llm=llm,
)
diff --git a/examples/features/multiple_tasks.py b/examples/features/multiple_tasks.py
deleted file mode 100644
index 0da02d666..000000000
--- a/examples/features/multiple_tasks.py
+++ /dev/null
@@ -1,84 +0,0 @@
-import asyncio
-import os
-import sys
-
-sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-
-from browser_use import Agent
-from browser_use.browser import BrowserProfile, BrowserSession
-from browser_use.browser.profile import ViewportSize
-from browser_use.llm import ChatGoogle
-
-api_key = os.getenv('GOOGLE_API_KEY')
-
-if not api_key:
- raise ValueError('GOOGLE_API_KEY is not set')
-
-llm = ChatGoogle(model='gemini-2.0-flash', api_key=api_key)
-
-
-async def main():
- # Create browser profile with settings
- browser_profile = BrowserProfile(
- headless=False,
- viewport=ViewportSize(width=1502, height=853),
- ignore_https_errors=True,
- )
-
- # Create browser session
- browser_session = BrowserSession(
- browser_profile=browser_profile,
- )
-
- # Start the browser
- await browser_session.start()
-
- agent = Agent(
- browser_session=browser_session,
- task='Go to https://browser-use.com/',
- llm=llm,
- )
-
- try:
- result = await agent.run()
- print(f'First task was {"successful" if result.is_successful else "not successful"}')
-
- if not result.is_successful:
- raise RuntimeError('Failed to navigate to the initial page.')
-
- agent.add_new_task('Navigate to the documentation page')
-
- result = await agent.run()
- print(f'Second task was {"successful" if result.is_successful else "not successful"}')
-
- if not result.is_successful:
- raise RuntimeError('Failed to navigate to the documentation page.')
-
- while True:
- next_task = input('Write your next task or leave empty to exit\n> ')
-
- if not next_task.strip():
- print('Exiting...')
- break
-
- agent.add_new_task(next_task)
- result = await agent.run()
-
- print(f"Task '{next_task}' was {'successful' if result.is_successful else 'not successful'}")
-
- if not result.is_successful:
- print('Failed to complete the task. Please try again.')
- continue
-
- finally:
- # Stop the browser session
- await browser_session.stop()
-
-
-if __name__ == '__main__':
- asyncio.run(main())
diff --git a/examples/features/outsource_state.py b/examples/features/outsource_state.py
deleted file mode 100644
index 0c286c5c4..000000000
--- a/examples/features/outsource_state.py
+++ /dev/null
@@ -1,64 +0,0 @@
-"""
-Show how to use custom outputs.
-
-@dev You need to add OPENAI_API_KEY to your environment variables.
-"""
-
-import asyncio
-import os
-import sys
-
-sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-import anyio
-
-from browser_use import Agent
-from browser_use.agent.views import AgentState
-from browser_use.browser import BrowserProfile, BrowserSession
-from browser_use.llm import ChatOpenAI
-
-
-async def main():
- task = 'Go to hackernews show hn and give me the first 5 posts'
-
- browser_profile = BrowserProfile(
- headless=True,
- )
- browser_session = BrowserSession(browser_profile=browser_profile)
-
- agent_state = AgentState()
-
- for i in range(10):
- agent = Agent(
- task=task,
- llm=ChatOpenAI(model='gpt-4.1-mini'),
- browser_session=browser_session,
- injected_agent_state=agent_state,
- page_extraction_llm=ChatOpenAI(model='gpt-4.1-mini'),
- )
-
- done, valid = await agent.take_step()
- print(f'Step {i}: Done: {done}, Valid: {valid}')
-
- if done and valid:
- break
-
- # Save state to file
- async with await anyio.open_file('agent_state.json', 'w') as f:
- serialized = agent_state.model_dump_json(exclude={'history'})
- await f.write(serialized)
-
- # Load state back from file
- async with await anyio.open_file('agent_state.json', 'r') as f:
- loaded_json = await f.read()
- agent_state = AgentState.model_validate_json(loaded_json)
-
- break
-
-
-if __name__ == '__main__':
- asyncio.run(main())
diff --git a/examples/features/parallel_agents.py b/examples/features/parallel_agents.py
index 58f3177d3..c2964ad27 100644
--- a/examples/features/parallel_agents.py
+++ b/examples/features/parallel_agents.py
@@ -23,6 +23,7 @@ browser_session = BrowserSession(
llm = ChatOpenAI(model='gpt-4.1-mini')
+# NOTE: This is experimental - you will have multiple agents running in the same browser session
async def main():
await browser_session.start()
agents = [
@@ -31,8 +32,8 @@ async def main():
'Search Google for weather in Tokyo',
'Check Reddit front page title',
'Look up Bitcoin price on Coinbase',
- 'Find NASA image of the day',
- 'Check top story on CNN',
+ # 'Find NASA image of the day',
+ # 'Check top story on CNN',
# 'Search latest SpaceX launch date',
# 'Look up population of Paris',
# 'Find current time in Sydney',
diff --git a/examples/features/pause_agent.py b/examples/features/pause_agent.py
deleted file mode 100644
index dc7148c99..000000000
--- a/examples/features/pause_agent.py
+++ /dev/null
@@ -1,101 +0,0 @@
-import asyncio
-import os
-import sys
-import threading
-
-sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-from browser_use import Agent, ChatOpenAI
-
-
-class AgentController:
- def __init__(self):
- llm = ChatOpenAI(model='gpt-4.1-mini')
- self.agent = Agent(
- task='open in one action https://www.google.com, https://www.wikipedia.org, https://www.youtube.com, https://www.github.com, https://amazon.com',
- llm=llm,
- )
- self.running = False
-
- async def run_agent(self):
- """Run the agent"""
- self.running = True
- await self.agent.run()
-
- def start(self):
- """Start the agent in a separate thread"""
- loop = asyncio.new_event_loop()
- asyncio.set_event_loop(loop)
- loop.run_until_complete(self.run_agent())
-
- def pause(self):
- """Pause the agent"""
- self.agent.pause()
-
- def resume(self):
- """Resume the agent"""
- self.agent.resume()
-
- def stop(self):
- """Stop the agent"""
- self.agent.stop()
- self.running = False
-
-
-def print_menu():
- print('\nAgent Control Menu:')
- print('1. Start')
- print('2. Pause')
- print('3. Resume')
- print('4. Stop')
- print('5. Exit')
-
-
-async def main():
- controller = AgentController()
- agent_thread = None
-
- while True:
- print_menu()
- try:
- choice = input('Enter your choice (1-5): ')
- except KeyboardInterrupt:
- choice = '5'
-
- if choice == '1' and not agent_thread:
- print('Starting agent...')
- agent_thread = threading.Thread(target=controller.start)
- agent_thread.start()
-
- elif choice == '2':
- print('Pausing agent...')
- controller.pause()
-
- elif choice == '3':
- print('Resuming agent...')
- controller.resume()
-
- elif choice == '4':
- print('Stopping agent...')
- controller.stop()
- if agent_thread:
- agent_thread.join()
- agent_thread = None
-
- elif choice == '5':
- print('Exiting...')
- if controller.running:
- controller.stop()
- if agent_thread:
- agent_thread.join()
- break
-
- await asyncio.sleep(0.1) # Small delay to prevent CPU spinning
-
-
-if __name__ == '__main__':
- asyncio.run(main())
diff --git a/examples/features/planner.py b/examples/features/planner.py
deleted file mode 100644
index 91c63ba78..000000000
--- a/examples/features/planner.py
+++ /dev/null
@@ -1,28 +0,0 @@
-import asyncio
-import os
-import sys
-
-sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-from browser_use import Agent, ChatOpenAI
-
-llm = ChatOpenAI(model='gpt-4.1', temperature=0.0)
-planner_llm = ChatOpenAI(
- model='o3-mini',
-)
-task = 'your task'
-
-
-agent = Agent(task=task, llm=llm, planner_llm=planner_llm, use_vision_for_planner=False, planner_interval=1)
-
-
-async def main():
- await agent.run()
-
-
-if __name__ == '__main__':
- asyncio.run(main())
diff --git a/examples/features/result_processing.py b/examples/features/process_agent_output.py
similarity index 100%
rename from examples/features/result_processing.py
rename to examples/features/process_agent_output.py
diff --git a/examples/features/restrict_urls.py b/examples/features/restrict_urls.py
index f995097f3..912fa5da7 100644
--- a/examples/features/restrict_urls.py
+++ b/examples/features/restrict_urls.py
@@ -12,7 +12,7 @@ from browser_use import Agent
from browser_use.browser import BrowserProfile, BrowserSession
from browser_use.llm import ChatOpenAI
-llm = ChatOpenAI(model='gpt-4.1', temperature=0.0)
+llm = ChatOpenAI(model='gpt-4.1-mini')
task = (
"go to google.com and search for openai.com and click on the first link then extract content and scroll down - what's there?"
)
diff --git a/examples/features/save_trace.py b/examples/features/save_trace.py
deleted file mode 100644
index 7dbe82375..000000000
--- a/examples/features/save_trace.py
+++ /dev/null
@@ -1,38 +0,0 @@
-import asyncio
-import os
-import sys
-
-sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-from browser_use.agent.service import Agent
-from browser_use.browser import BrowserProfile, BrowserSession
-from browser_use.llm import ChatOpenAI
-
-llm = ChatOpenAI(model='gpt-4.1', temperature=0.0)
-
-
-async def main():
- browser_session = BrowserSession(
- browser_profile=BrowserProfile(
- traces_dir='./tmp/traces/',
- user_data_dir='~/.config/browseruse/profiles/default',
- )
- )
-
- await browser_session.start()
- try:
- agent = Agent(
- task='Go to hackernews, then go to apple.com and return all titles of open tabs',
- llm=llm,
- browser_session=browser_session,
- )
- await agent.run()
- finally:
- await browser_session.stop()
-
-
-asyncio.run(main())
diff --git a/examples/features/small_model_for_extraction.py b/examples/features/small_model_for_extraction.py
index 6d44a2ff2..3b18c0e77 100644
--- a/examples/features/small_model_for_extraction.py
+++ b/examples/features/small_model_for_extraction.py
@@ -10,8 +10,11 @@ load_dotenv()
from browser_use import Agent, ChatOpenAI
-llm = ChatOpenAI(model='gpt-4.1', temperature=0.0)
-small_llm = ChatOpenAI(model='gpt-4.1-mini', temperature=0.0)
+# This uses a bigger model for the planning
+# And a smaller model for the page content extraction
+# THink of it like a subagent which only task is to extract content from the current page
+llm = ChatOpenAI(model='gpt-4.1')
+small_llm = ChatOpenAI(model='gpt-4.1-mini')
task = 'Find the founders of browser-use in ycombinator, extract all links and open the links one by one'
agent = Agent(task=task, llm=llm, page_extraction_llm=small_llm)
diff --git a/examples/features/validate_output.py b/examples/features/validate_output.py
deleted file mode 100644
index 31ffed812..000000000
--- a/examples/features/validate_output.py
+++ /dev/null
@@ -1,49 +0,0 @@
-"""
-Demonstrate output validator.
-
-@dev You need to add OPENAI_API_KEY to your environment variables.
-"""
-
-import asyncio
-import os
-import sys
-
-sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
-
-from dotenv import load_dotenv
-
-load_dotenv()
-
-from pydantic import BaseModel
-
-from browser_use import ActionResult, Agent, Controller
-from browser_use.llm import ChatOpenAI
-
-controller = Controller()
-
-
-class DoneResult(BaseModel):
- title: str
- comments: str
- hours_since_start: int
-
-
-# we overwrite done() in this example to demonstrate the validator
[email protected]('Done with task', param_model=DoneResult)
-async def done(params: DoneResult):
- result = ActionResult(is_done=True, extracted_content=params.model_dump_json())
- print(result)
- # NOTE: this is clearly wrong - to demonstrate the validator
- return 'blablabla'
-
-
-async def main():
- task = 'Go to hackernews hn and give me the top 1 post'
- model = ChatOpenAI(model='gpt-4.1-mini')
- agent = Agent(task=task, llm=model, controller=controller, validate_output=True)
- # NOTE: this should fail to demonstrate the validator
- await agent.run(max_steps=5)
-
-
-if __name__ == '__main__':
- asyncio.run(main())
diff --git a/examples/file_system/excel_sheet.py b/examples/file_system/excel_sheet.py
index e0164c998..3f80cfe93 100644
--- a/examples/file_system/excel_sheet.py
+++ b/examples/file_system/excel_sheet.py
@@ -9,20 +9,10 @@ sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from dotenv import load_dotenv
load_dotenv()
-from lmnr import Laminar
-
-try:
- Laminar.initialize(project_api_key=os.getenv('LMNR_PROJECT_API_KEY'))
-except Exception:
- pass
from browser_use import Agent
-# Initialize the model
-llm = ChatOpenAI(
- model='o4-mini',
- temperature=1.0,
-)
+llm = ChatOpenAI(model='o4-mini')
task = (