Literature Review Agent

 

AWS Marketplace

The quickest way to get up and running is through the AWS Marketplace. A few clicks and the full appliance is deployed on an EC2 instance in your own AWS account — nothing leaves your environment.

Visit the product page on AWS Marketplace and follow the steps below.

Getting started:

  1. Subscribe to the product on the AWS Marketplace.

  2. Launch the AMI on a new EC2 instance from the Marketplace console.
    • We recommend m5.2xlarge or larger (8 vCPU / 32 GB RAM) for most workloads.
    • Only port 80 needs to be open in the security group — the UI runs on port 80.

  3. Wait for the services to come up. The first boot pulls Docker images and initialises the database. Expect around 10–15 minutes, depending on your region. You’ll see a loading screen until everything is ready.

  4. Open the workspace in your browser:
    • Navigate to http://<INSTANCE_PUBLIC_IP>.
    • The default password is the EC2 instance ID — you’ll find it in the AWS EC2 console.
  5. Run the setup wizard (/setup):
    • Pick your LLM provider: Amazon SageMaker (models running in your own account) or the DeepLens API (api.deeplens.health).
    • Enter your API key.
    • Optionally configure keys for external search sources — NCBI, Semantic Scholar, Europe PMC, and OpenAlex are all supported but none are required.
Use Instance type vCPU RAM
Evaluation / small teams m5.xlarge 4 16 GB
Standard workloads m5.2xlarge 8 32 GB
Large / parallel runs m5.4xlarge 16 64 GB

Configuration

All configuration is managed through the in-app setup wizard. On first boot, the wizard walks you through connecting your LLM and search API keys — no manual file editing required.

API keys configured through the setup wizard:

Key Purpose
DeepLens API key LLM and search access (required)
NCBI / PubMed API key Higher rate limits for PubMed searches (optional)
Semantic Scholar API key Higher rate limits (optional)
Europe PMC API key Higher rate limits (optional)
OpenAlex API key Higher rate limits (optional)

Secure HTTPS Access via CloudFront

Out of the box, the appliance listens on HTTP port 80. If you need HTTPS access from the public internet, the simplest option is to put a CloudFront distribution in front of your EC2 instance. The CloudFront resources involved are Free Tier Eligible, and the setup takes only a few minutes.

  1. Create the CloudFormation stack YAML:

    AWSTemplateFormatVersion: '2010-09-09'
Description: >
  Expose Literature Review over HTTPS via CloudFront.
  Provide the EC2 instance hostname as input.
Parameters:
  LrInstanceHostName:
    Description: Public hostname or IP of the Literature Review EC2 instance
    Type: String
Resources:
  CloudFront:
    Type: AWS::CloudFront::Distribution
    Properties:
      DistributionConfig:
        Enabled: true
        DefaultCacheBehavior:
          AllowedMethods: [DELETE, GET, HEAD, OPTIONS, PATCH, POST, PUT]
          DefaultTTL: 0
          MaxTTL: 0
          ForwardedValues:
            QueryString: true
            Cookies:
              Forward: all
          TargetOriginId: LrOrigin
          ViewerProtocolPolicy: redirect-to-https
        Origins:
          - Id: LrOrigin
            DomainName: !Ref LrInstanceHostName
            CustomOriginConfig:
              HTTPPort: 80
              OriginProtocolPolicy: http-only
  2. Deploy the stack through the AWS Console or CLI, passing in your EC2 instance’s public hostname.
  3. Once the stack is deployed, access Literature Review Agent using the CloudFront distribution domain name over HTTPS.
Last updated