Awesome Prompts for Browser-Use

To achieve high-quality results with Browser-Use, follow these prompt engineering principles:

Structure

• Be specific about the goal: Clearly state the objective.
• Break complex tasks into steps: Use numbered lists for multi-step processes.
• Include format specifications: Mention desired output formats (e.g., JSON, Markdown, Google Docs).
• Provide context: Include relevant background information.
• Set constraints: Specify limitations or requirements.

Clarity

• Use precise language: Avoid ambiguity.
• Specify parameters: Include exact values, ranges, or criteria.
• Clarify expectations: Explain what success looks like.
• Include examples: Provide examples of desired outputs.
• Define terminology: Explain industry-specific or technical terms.

Adaptability

• Include fallback options: Provide alternatives if the primary approach fails.
• Specify error handling: Explain how to deal with potential issues.
• Allow for iteration: Build in opportunities for refinement.
• Request verification: Ask the agent to verify critical information.
• Include decision points: Define how choices should be made when alternatives exist.

Efficiency

• Prioritize tasks: Indicate which aspects are most important.
• Set time or resource limits: Specify constraints on depth or breadth.
• Focus on value-add activities: Emphasize where the agent provides the most value.
• Minimize unnecessary steps: Avoid redundant activities.
• Batch similar operations: Group related tasks for efficiency.

The repository provides structured, multi-step prompts designed for complex, long-running workflows. These prompts use a phased approach (e.g., Research Phase, Analysis Phase, Execution Phase) to guide an agent through high-level objectives.

Key workflow patterns include:

• Research and Report Generation: Automates market research, data organization in Google Sheets, and professional report creation in Google Docs.
• Event Planning: Manages venue research, vendor selection, budget allocation, and timeline creation.
• Website Migration: Handles content auditing, technical planning, migration timelines, and testing/launch strategies.
• Product Launch Campaigns: Covers market research, strategy development, content planning, and post-launch measurement.

### Research and Report Generation Example

```text
Create a comprehensive market research report on {industry} trends and competitive landscape.

1. Research Phase:
   - Search for industry reports from sources like Statista, IBISWorld, and Gartner
   - Find recent news articles about {industry} from the past 6 months
   - Identify the top 5 companies in this space and gather key metrics
   - Locate market size estimates and growth projections
   - Find information about emerging technologies or disruptions in this sector

2. Data Organization Phase:
   - Create a Google Sheet to compile numerical data and key statistics
   - Organize company profiles with strengths, weaknesses, and unique offerings
   - Track market share percentages and changes over time
   - Compile consumer trend data and demographic information

3. Analysis Phase:
   - Identify 3-5 key trends shaping the industry
   - Analyze competitive positioning of major players
   - Evaluate market opportunities and threats
   - Examine regulatory factors affecting the industry
   - Assess technological impacts on future growth

4. Report Creation Phase:
   - Create a Google Doc with a professional structure including:
     - Executive summary
     - Industry overview with market size and growth rates
     - Competitive landscape analysis
     - Detailed trend analysis with supporting data
     - Future outlook and strategic recommendations
     - References and methodology section
   - Add appropriate charts and tables from your research
   - Format the document professionally with consistent styling

The final report should be 10-15 pages, include proper citations, and be suitable for presentation to executive stakeholders.

Use these prompt templates to automate web-based research tasks using Browser-Use agents. They include structured workflows for:

• Comparative Research: Comparing pricing and features of competitors in a specific industry.
• Academic Research: Finding and extracting metadata (title, authors, DOI, citations) from recent research papers, including handling paywalls by looking for preprints.
• News Summary: Synthesizing developments on a topic from multiple reputable sources, including headline extraction and perspective analysis.
• Fact Verification: Verifying claims by cross-referencing multiple authoritative sources and determining accuracy levels (Accurate, Partially accurate, Inaccurate, or Unverifiable).

### Comparative Research Example

Research and compare the pricing and key features of the top 3 competitors in the {industry} space: {competitor1}, {competitor2}, and {competitor3}.

For each competitor:

1. Navigate to their pricing page
2. Extract the pricing tiers and their costs
3. List the key features in each tier
4. Note any special offers or discounts currently available

Present the information in a structured format that makes it easy to compare the options.

Leverage Browser-Use agents to create content directly within Google Workspace applications:

• Blog Post Draft: Researching topics on Google and drafting structured posts in Google Docs.
• Social Media Content Calendar: Creating calendars in Google Sheets with themes, copy, hashtags, and posting schedules.
• Email Newsletter: Drafting newsletters in email platforms with specific sections like subject lines, feature articles, and CTAs.
• Presentation Creation: Building professional slide decks in Google Slides following specific rules like the 6x6 rule and including speaker notes.

### Blog Post Draft Example

Go to Google Docs and create a new document. Write a well-structured blog post about {topic} following these steps:

1. Research the top 5 articles on Google about {topic} to understand current perspectives
2. Draft a compelling headline that includes the primary keyword
3. Write an introduction that highlights the key problem or opportunity
4. Create 3-5 main sections with appropriate subheadings
5. Include bullet points or numbered lists where appropriate
6. Add a conclusion with a clear call-to-action
7. Suggest 2-3 relevant images that could be included (with placeholder descriptions)

The blog post should be approximately 1000-1200 words and written in a {formal/conversational/educational} tone.

Automate e-commerce workflows with these prompt templates:

• Product Price Comparison: Finding the best price for a product across major retailers (Amazon, Walmart, etc.), including shipping and stock status.
• Shopping Cart Automation: Adding specific items with defined specifications and quantities to a cart, handling alternatives if items are unavailable, and stopping before payment.
• Deal Finder: Searching for deals in a specific category, filtering by discount percentage, and verifying customer ratings.
• Product Research: Conducting deep research into models/brands based on price ranges, required features, and expert/customer reviews.

### Shopping Cart Automation Example

Add the following items to my shopping cart on {website}:

1. {item1} - {specifications1} - quantity: {quantity1}
2. {item2} - {specifications2} - quantity: {quantity2}
3. {item3} - {specifications3} - quantity: {quantity3}

For each item:

• If the exact item is unavailable, find the closest alternative
• If there are multiple options (size, color, etc.), select {preference}
• If the price is above {budget} for any item, look for a more affordable alternative

Once all items are in the cart:

• Apply coupon code: {coupon_code} if available
• Select the standard shipping option
• Proceed to checkout but stop before payment

Extract specific data types from websites using these templates:

• Tabular Data Extraction: Navigating to specific pages or search queries to extract rows/columns from tables into CSV-compatible formats, including pagination handling.
• Contact Information Extraction: Gathering office addresses, phone numbers, emails, and executive details, cross-referencing with LinkedIn or Google Business profiles.
• Product Specifications Extraction: Collecting technical, performance, and compatibility data from manufacturers and retailers.
• Competitor Analysis Extraction: Gathering mission statements, product lines, pricing, and marketing strategies for top competitors.

### Tabular Data Extraction Example

Extract structured data from {website} about {data_category}.

1. Navigate to {specific_page_or_search_query}
2. Identify the table or structured data containing information about {data_points}
3. Extract all rows and columns while preserving the relationship between data points
4. If pagination exists, navigate through at least the first 3 pages and extract all data
5. If filters are available, apply {filter_criteria} before extraction

Format the extracted data in a CSV-compatible format with appropriate headers. If any data points are missing, mark them as "N/A" rather than leaving them blank.

This prompt enables brand monitoring across multiple platforms. The agent searches for brand mentions, categorizes them (positive, customer service, product questions, etc.), identifies high-impact conversations, and drafts appropriate responses based on brand guidelines.

Conduct social listening for {brand_name} across {platform1}, {platform2}, and {platform3} and help craft appropriate responses.

1. Search for recent mentions of {brand_name} and related terms {related_term1}, {related_term2}
2. Categorize mentions as:
   - Positive feedback
   - Customer service issues
   - Product questions
   - General mentions
   - Competitor comparisons

3. For each category, identify the 3-5 most important conversations based on:
   - User influence/follower count
   - Engagement level on the post
   - Sentiment intensity
   - Potential impact on brand reputation

4. For each identified conversation, draft an appropriate response following our response guidelines:
   - Positive feedback: Thank them and reinforce brand relationship
   - Customer service: Acknowledge issue, express empathy, move to private channel
   - Product questions: Provide accurate information with links to resources
   - General mentions: Engage authentically in alignment with brand voice
   - Competitor comparisons: Highlight differentiators without disparaging competitors

This prompt automates the process of filling out tax forms by extracting data from various document types (W2, 1099, deduction docs, etc.). It is designed to prepare a draft without submitting the final return, allowing for human verification.

Help me complete my {tax_form_type} for {tax_year}.

1. Navigate to {tax_website} and log in with the provided credentials
2. Begin a new {tax_form_type} for {tax_year}
3. For each section, enter the appropriate information from my documents:
   - Personal information from {personal_doc}
   - Income information from {w2_docs}, {1099_docs}
   - Deduction information from {deduction_docs}
   - Credit information from {credit_docs}
   - Investment information from {investment_docs}

4. For any sections where information is unclear:
   - Flag the section for my review
   - Explain what additional information is needed
   - Suggest potential sources for the information

5. Before submitting:
   - Review all entries for accuracy
   - Compare key figures with last year's return for significant discrepancies
   - Calculate the estimated refund or amount owed
   - Save the return as a draft for my final review

Do not submit the final return, but prepare it to the point where it is ready for my final verification and submission.

Use this prompt to find potential brand partners. The agent searches for influencers within specific follower ranges and niches, analyzes their content quality, engagement rates, and audience sentiment, and produces a ranked shortlist of candidates.

Research potential influencers for a partnership with {brand_name} to promote {product/service}.

1. On {platform1} and {platform2}, find 10 influencers who:
   - Have between {min_followers} and {max_followers} followers
   - Create content related to {niche/industry}
   - Have an audience demographic matching {target_audience}
   - Have an engagement rate of at least {min_engagement_rate}%

2. For each influencer, analyze:
   - Content quality and brand alignment
   - Posting frequency and consistency
   - Previous brand partnerships
   - Audience sentiment in comments
   - Any potential red flags or controversies

3. Create a ranked shortlist of the top 5 candidates with:
   - Profile links
   - Key audience demographics
   - Estimated reach and engagement metrics
   - Suggested partnership approach (sponsored post, takeover, affiliate, etc.)
   - Estimated appropriate compensation range based on industry standards

Automate customer feedback surveys. The prompt handles rating scales (1-5, 1-10), multiple-choice questions, and open-ended qualitative responses based on provided sentiment and feedback data.

Complete the customer feedback survey for {product/service} from {company}.

1. Navigate to the survey link: {survey_link}
2. For rating questions (1-5 or 1-10 scales):
   - Overall satisfaction: {overall_rating}
   - Product quality: {quality_rating}
   - Customer service: {service_rating}
   - Value for money: {value_rating}
   - Likelihood to recommend: {recommendation_rating}

3. For multiple-choice questions, select options that align with:
   - Frequency of use: {usage_frequency}
   - Primary purpose: {usage_purpose}
   - Features used most often: {key_features}

4. For open-ended questions, write thoughtful responses about:
   - What I liked most: {positive_aspects}
   - What could be improved: {improvement_suggestions}
   - Additional features desired: {desired_features}

5. Demographic questions (if asked):
   - Fill in basic information without sharing unnecessarily sensitive details
   - Skip optional demographic questions unless specifically instructed to complete them

Complete the survey honestly but constructively, providing feedback that would be helpful for product improvement.

This prompt directs an agent to log into LinkedIn and perform a comprehensive audit of a user's profile. It includes analyzing the headline, about section, experience, and skills, as well as researching successful professionals in the same field to identify areas for improvement.

Help me optimize my LinkedIn profile to improve my chances of getting noticed for {target_job_title} roles.

1. Navigate to LinkedIn and log in with my credentials
2. Go to my profile page
3. Analyze each section of my profile and suggest improvements:
   - Profile photo and banner
   - Headline (create one that includes key terms for my target role)
   - About section (craft a compelling summary highlighting relevant skills)
   - Experience section (enhance job descriptions with accomplishments and metrics)
   - Skills section (suggest the top 10 skills I should list based on {target_job_title})
   - Education and certifications
   - Recommendations and endorsements

4. Research profiles of 3 successful professionals in similar roles and identify elements I should incorporate
5. Suggest 5 LinkedIn groups I should join related to {industry}
6. Draft 3 potential posts I could share to demonstrate expertise in my field