Browser Use excels at gathering information from multiple websites, analyzing content, and synthesizing research findings.
Basic Research Task
Gather information on a specific topic:
import asyncio
from browser_use import Agent, ChatBrowserUse
async def main():
task = """
Research the latest developments in quantum computing:
1. Search for recent news articles (last 30 days)
2. Find 3-5 key developments or breakthroughs
3. Summarize each finding with:
- Title
- Source
- Date
- Key points
4. Identify common themes across articles
"""
agent = Agent(task=task, llm=ChatBrowserUse())
result = await agent.run()
if __name__ == '__main__':
asyncio.run(main())
Multi-Source Research
Gather information from specific sources:
task = """
Research electric vehicle market trends:
Sources to check:
1. Tesla investor relations page
2. Bloomberg automotive section
3. TechCrunch transportation category
For each source, extract:
- Latest articles/reports about EV sales
- Market share data
- Future predictions
Compile findings into a comprehensive summary comparing
perspectives from each source.
"""
agent = Agent(
task=task,
llm=ChatBrowserUse(),
max_steps=100 # Allow more steps for thorough research
)
import asyncio
from pydantic import BaseModel
from browser_use import Agent, ChatOpenAI, Tools
from browser_use.agent.views import ActionResult
class Profile(BaseModel):
platform: str
profile_url: str
class Profiles(BaseModel):
profiles: list[Profile]
tools = Tools(exclude_actions=['search'], output_model=Profiles)
@tools.registry.action('Search the web for a specific query')
async def search_web(query: str):
"""Custom search implementation for better results"""
# Your search API implementation here
pass
async def main():
task = """
Find social media profiles for the TikTok user from this video:
https://www.tiktok.com/share/video/7470981717659110678/
1. Open the TikTok video
2. Extract the @username from the URL or page
3. Search the web for this username
4. Find their profiles on:
- Instagram
- Twitter/X
- YouTube
- LinkedIn
5. Return the profile URLs with platform names
"""
agent = Agent(task=task, llm=ChatOpenAI(model='gpt-4.1-mini'), tools=tools)
history = await agent.run()
# Access structured output
if history and history.structured_output:
profiles = history.structured_output
for profile in profiles.profiles:
print(f'{profile.platform}: {profile.profile_url}')
if __name__ == '__main__':
asyncio.run(main())
Competitive Analysis
Compare competitors across multiple dimensions:
from pydantic import BaseModel, Field
class CompanyInfo(BaseModel):
name: str
website: str
pricing_model: str
key_features: list[str]
target_market: str
latest_news: str
class CompetitiveAnalysis(BaseModel):
competitors: list[CompanyInfo]
market_leader: str
key_differentiators: dict[str, str]
async def analyze_competitors(industry: str):
task = f"""
Research top 5 companies in the {industry} industry:
For each company:
1. Visit their website
2. Extract:
- Company name
- Pricing model
- Key features/products
- Target market
- Recent news or product launches
Analysis:
- Identify the market leader
- Compare key differentiators
- Note unique selling propositions
"""
agent = Agent(
task=task,
llm=ChatBrowserUse(),
output_model_schema=CompetitiveAnalysis
)
return await agent.run()
News Aggregation
Collect news from multiple sources on a topic:
task = """
Aggregate news about artificial intelligence from:
1. TechCrunch (https://techcrunch.com)
2. The Verge (https://www.theverge.com)
3. Hacker News (https://news.ycombinator.com)
For each site:
- Find the top 3 AI-related articles from today
- Extract: headline, summary, publication time, URL
Organize by publication time (newest first) and
identify any overlapping stories across sources.
"""
agent = Agent(task=task, llm=ChatBrowserUse())
Academic Research
Gather information from academic sources:
task = """
Research papers on machine learning optimization:
1. Go to Google Scholar (https://scholar.google.com)
2. Search for "machine learning optimization techniques"
published in last 2 years
3. Extract top 10 papers with:
- Title
- Authors
- Publication venue
- Citation count
- Abstract (first 200 words)
- PDF link if available
4. Identify the most cited papers
5. Find common methodologies across papers
"""
agent = Agent(
task=task,
llm=ChatBrowserUse(),
max_steps=80
)
Product Research
Research product specifications and reviews:
task = """
Research the iPhone 15 Pro:
Information to gather:
1. Official specs from Apple.com:
- Display size and technology
- Processor
- Camera specifications
- Battery life
- Storage options
- Price for each variant
2. Reviews from:
- CNET
- The Verge
- TechRadar
3. For each review, extract:
- Overall rating
- Pros and cons
- Verdict summary
4. Compare findings and create a comprehensive summary
highlighting consensus opinions and outliers.
"""
Market Research
Analyze market trends and consumer sentiment:
task = """
Market research for sustainable fashion brands:
1. Identify top 10 sustainable fashion brands
2. For each brand visit their website and extract:
- Sustainability practices
- Price range
- Target demographic
- Unique value proposition
3. Check social media presence:
- Instagram follower count
- Recent post engagement
- Brand sentiment in comments
4. Search for industry reports on sustainable fashion market size
5. Compile findings into market overview with:
- Market leaders
- Price positioning
- Growth trends
- Consumer preferences
"""
agent = Agent(
task=task,
llm=ChatBrowserUse(),
max_steps=150 # Complex research needs more steps
)
Wikipedia Research
Extract structured information from Wikipedia:
task = """
Research quantum computing on Wikipedia:
1. Go to Wikipedia article on Quantum Computing
2. Extract:
- Introduction summary
- Key principles (from "Principles" section)
- Current applications
- Major companies/organizations involved
- Timeline of developments
- References to notable papers
3. Follow links to related articles:
- Quantum supremacy
- Quantum algorithm
4. Create a comprehensive overview with sources
"""
Research Best Practices
Define Clear Objectives
Specify exactly what information you need and from which sources
Structure Your Output
Use Pydantic models to ensure consistent data format across sources
Verify Information
Cross-reference facts from multiple sources for accuracy
Save Your Progress
Use save_conversation_path to preserve research history
Performance Tip: For extensive research, increase max_steps to allow thorough investigation across multiple sites.
Always respect website terms of service and robots.txt when conducting research at scale.
Advanced Research Patterns
Iterative Deep Dive
task = """
Start with broad topic "renewable energy", then:
1. Identify the 3 most promising subtopics
2. For each subtopic, find top 3 authoritative sources
3. From those sources, extract key statistics and trends
4. Synthesize findings into a structured report
"""
Comparative Analysis
task = """
Compare cloud providers (AWS vs Azure vs GCP):
1. Visit pricing pages for each
2. Compare equivalent services (compute, storage, database)
3. Check documentation quality
4. Review customer case studies
5. Create side-by-side comparison table
"""