Simple Research Crew

A basic research crew example with researcher and writer agents

Simple Research Crew

This example demonstrates how to create a basic research crew with two agents: a researcher who gathers information and a writer who creates content based on that research.

Overview

We’ll create a crew that:

Researches a given topic
Writes an informative article based on the research
Returns both the research findings and the final article

Complete Code

require 'rcrewai'

# Configure RCrewAI - supports multiple providers
RCrewAI.configure do |config|
  config.llm_provider = :openai  # or :anthropic, :google, :azure
  # API keys are loaded automatically from environment variables
  config.openai_model = 'gpt-4'  # or 'gpt-3.5-turbo' for lower cost
  config.temperature = 0.1       # Lower = more consistent, Higher = more creative
end

# Create the crew
research_crew = RCrewAI::Crew.new("research_crew")

# Define the research agent
researcher = RCrewAI::Agent.new(
  name: "senior_researcher",
  role: "Senior Research Analyst",
  goal: "Uncover cutting-edge developments in AI and technology",
  backstory: "You work at a leading tech think tank. Your expertise lies in identifying emerging trends in AI and technology. You have a knack for dissecting complex topics and presenting clear insights.",
  verbose: true,
  allow_delegation: false
)

# Define the writer agent
writer = RCrewAI::Agent.new(
  name: "tech_writer",
  role: "Tech Content Strategist", 
  goal: "Craft compelling content on tech advancements",
  backstory: "You are a renowned Content Strategist, known for your insightful and engaging articles on technology and innovation. With a deep understanding of the tech industry, you transform complex concepts into compelling narratives.",
  verbose: true,
  allow_delegation: true
)

# Add agents to the crew
research_crew.add_agent(researcher)
research_crew.add_agent(writer)

# Define the research task
research_task = RCrewAI::Task.new(
  name: "research_ai_advancements",
  description: "Conduct a comprehensive analysis of the latest advancements in AI in 2024. Identify key trends, breakthrough technologies, and their potential impact on various industries. Your final report should be detailed and well-structured.",
  agent: researcher,
  expected_output: "A comprehensive research report with key findings, trends, and implications of AI advancements in 2024"
)

# Define the writing task
writing_task = RCrewAI::Task.new(
  name: "write_ai_article",
  description: "Using the research provided, write a compelling blog article about the latest AI advancements. Make it engaging but informative, suitable for a tech-savvy audience. Include an introduction, main content with key points, and a conclusion.",
  agent: writer,
  expected_output: "A well-written, engaging blog article about AI advancements (800-1000 words)",
  context: [research_task]  # This task depends on the research task
)

# Add tasks to the crew
research_crew.add_task(research_task)
research_crew.add_task(writing_task)

# Execute the crew
puts "Starting research crew execution..."
results = research_crew.execute

# Display results
puts "\n" + "="*50
puts "RESEARCH RESULTS:"
puts "="*50
puts research_task.result

puts "\n" + "="*50  
puts "FINAL ARTICLE:"
puts "="*50
puts writing_task.result

# Optional: Save results to files
File.write("research_findings.md", research_task.result)
File.write("ai_article.md", writing_task.result)

puts "\nResults saved to research_findings.md and ai_article.md"

Step-by-Step Explanation

1. Setup and Configuration

require 'rcrewai'

RCrewAI.configure do |config|
  config.llm_provider = :openai
  config.api_key = ENV['OPENAI_API_KEY']
  config.model = 'gpt-4'
end

First, we require the RCrewAI gem and configure it with our LLM provider. The API key should be stored in an environment variable for security.

2. Create the Crew

research_crew = RCrewAI::Crew.new("research_crew")

We create a new crew instance that will coordinate our agents and tasks.

3. Define Agents

Research Agent

researcher = RCrewAI::Agent.new(
  name: "senior_researcher",
  role: "Senior Research Analyst", 
  goal: "Uncover cutting-edge developments in AI and technology",
  backstory: "You work at a leading tech think tank...",
  verbose: true,
  allow_delegation: false
)

The researcher agent is designed to gather and analyze information. Key attributes:

Role: Defines what the agent does
Goal: Specific objective for the agent
Backstory: Provides context that influences the agent’s behavior
verbose: Enables detailed logging
allow_delegation: Controls whether this agent can delegate tasks

Writer Agent

writer = RCrewAI::Agent.new(
  name: "tech_writer",
  role: "Tech Content Strategist",
  goal: "Craft compelling content on tech advancements", 
  backstory: "You are a renowned Content Strategist...",
  verbose: true,
  allow_delegation: true
)

The writer transforms research into engaging content.

4. Define Tasks

Research Task

research_task = RCrewAI::Task.new(
  name: "research_ai_advancements",
  description: "Conduct a comprehensive analysis...",
  agent: researcher,
  expected_output: "A comprehensive research report..."
)

Tasks define what work needs to be done. The expected_output helps guide the agent’s work quality.

Writing Task

writing_task = RCrewAI::Task.new(
  name: "write_ai_article", 
  description: "Using the research provided...",
  agent: writer,
  expected_output: "A well-written, engaging blog article...",
  context: [research_task]  # Dependencies
)

The context parameter creates dependencies - the writing task will have access to the research task’s results.

5. Execute and Handle Results

results = research_crew.execute

puts research_task.result
puts writing_task.result

The crew executes tasks in order, respecting dependencies. Results are accessible through each task’s result attribute.

Running the Example

Install dependencies:
```
bundle install
```

Set environment variables (choose your provider):

# OpenAI
export OPENAI_API_KEY="your-openai-key"
   
# Or Anthropic
export ANTHROPIC_API_KEY="your-anthropic-key"
   
# Or Google
export GOOGLE_API_KEY="your-google-key"
   
# Or Azure
export AZURE_OPENAI_API_KEY="your-azure-key"

Run the script:
```
ruby research_crew_example.rb
```

Expected Output

The script will produce:

Research findings about AI advancements
A well-structured blog article based on the research
Saved files with the results

Customization Options

Different Topics

Change the task descriptions to research different topics:

research_task = RCrewAI::Task.new(
  name: "research_blockchain",
  description: "Research the latest developments in blockchain technology...",
  # ... rest of configuration
)

Adding Tools

Equip agents with tools for enhanced capabilities:

web_search_tool = RCrewAI::Tools::WebSearch.new

researcher = RCrewAI::Agent.new(
  name: "senior_researcher",
  # ... other config
  tools: [web_search_tool]
)

Multiple Output Formats

Create additional tasks for different output formats:

# Add a summarizer agent
summarizer = RCrewAI::Agent.new(
  name: "content_summarizer",
  role: "Content Summarizer",
  goal: "Create concise summaries"
)

# Summary task
summary_task = RCrewAI::Task.new(
  name: "create_summary",
  description: "Create a 200-word summary of the research and article",
  agent: summarizer,
  context: [research_task, writing_task]
)

Troubleshooting

Common Issues

API Key Issues: Ensure your OpenAI API key is correctly set
Agent Confusion: Make sure agent roles and goals are specific and clear
Task Dependencies: Verify that dependent tasks are added to the crew before dependent tasks

Debugging Tips

Set verbose: true on agents to see detailed execution logs
Use crew.verbose = true for crew-level logging
Check individual task results with task.result after execution

Next Steps

Try modifying the agent roles and backstories
Add more agents for specialized tasks (fact-checker, SEO optimizer)
Experiment with different task dependencies
Add custom tools for specific capabilities