A multi-agent research system using LangGraph for automated research and report generation

Build, test, and deploy a multi-agent AI system using LangGraph, Docker, AWS Lambda, and CircleCI. The system uses a research-driven AI workflow where different agents,such as fact-checking, summarization, and search agents, work together seamlessly. This application is packaged into a Docker container, deployed to AWS Lambda, and the entire pipeline is run using CircleCI.

The project has been developed as part of the following blog

• Features
• Prerequisites
• Installation
• Usage
  • Configuration
  • Local Execution
  • AWS Lambda Deployment
  • AWS Lambda Invocation
• License

• Multi-agent architecture using LangGraph
• Automated web search using Serper API
• Fact-checking and verification
• Report generation with structured summaries
• AWS Lambda deployment support
• Configurable confidence scores and retry mechanisms

• Python 3.12
• AWS CLI (for Lambda deployment)
• Serper API key
• OpenAI API key
• AWS Credentials (for Lambda deployment)

1. Clone the repository:

   git clone https://github.com/benitomartin/multiagent-langgraph-circleci.git
   cd multiagent-langgraph-circleci

2. Create a virtual environment:

   uv venv

3. Activate the virtual environment:

   • On Windows:

     .venv\Scripts\activate

   • On Unix or MacOS:

     source .venv/bin/activate

4. Install the required packages:

   uv sync --all-extras

5. Create a .env file in the root directory:

   # API Keys
   SERPER_API_KEY=your_serper_key_here                
   OPENAI_API_KEY=your_openai_key_here                
   
   # AWS Configuration
   AWS_REGION=your_aws_region                          
   AWS_ACCESS_KEY_ID=your_aws_access_key              
   AWS_SECRET_ACCESS_KEY=your_aws_secret_key          
   AWS_ACCOUNT_ID=your_aws_account_id                 
   
   # Repository and Image Configuration
   REPOSITORY_NAME=langgraph-ecr-docker-repo          
   IMAGE_NAME=langgraph-lambda-image                  
   
   # Lambda Configuration
   LAMBDA_FUNCTION_NAME=langgraph-lambda-function     
   ROLE_NAME=lambda-bedrock-role                      
   ROLE_POLICY_NAME=LambdaBedrockPolicy              
   

   To obtain the required API keys:

   • Serper API Key: Sign up at Serper.dev
   • OpenAI API Key: Sign up at OpenAI Platform
   • AWS Credentials: Create through AWS IAM Console

The following parameters can be adjusted in config/settings.py:

• CONFIDENCE_THRESHOLD: Threshold for confidence in fact-checking
• MAX_RETRIES: Maximum number of retries for the search agent
• ADD_MAX_RESULTS: Number of search results to add in each retry
• FACT_CHECK_MODEL: Model used for fact-checking (default: "gpt-4-mini")
• SUMMARIZATION_MODEL: Model used for summarization (default: "anthropic.claude-3-haiku")

To run the research graph locally:

uv run src/graph/research_graph.py \
   --query "What are the benefits of using AWS Cloud Services?" \
   --confidence-threshold 0.85 \
   --max-retries 3 \
   --add-max-results 2

Build and deploy the Docker image with the lambda function:

chmod +x build_deploy.sh
./build_deploy.sh

To invoke the deployed Lambda function add your region and run the following command:

aws lambda invoke \
    --function-name langgraph-lambda-function \
    --payload '{"query": "What are the benefits of using CircleCI?"}' \
    --region <your_region> \
    --cli-binary-format raw-in-base64-out \
    response.json && \
    cat response.json | jq

This project is licensed under the MIT License - see the LICENSE file for details.