Literature Review Agent

The Problem:

Systematic and scoping reviews are slow and repetitive. Teams spend weeks searching multiple databases, de-duplicating results, screening titles and abstracts against inclusion/exclusion criteria, and hand-extracting the same data points (study design, sample size, outcomes, etc. ) into spreadsheets. The work is error-prone, hard to reproduce, and difficult to audit.

The Solution:

Literature Review Agent is a guided workspace that automates the heavy lifting of evidence synthesis while keeping a human in control. You search trusted medical and academic sources, define what you want to extract in plain language, and let medical-grade LLMs pull structured, evidence-backed data from every article - with inclusion/exclusion screening applied automatically. Results are reviewable in the UI and exportable to CSV.

Literature Review Agent runs as a self-contained appliance inside your own cloud account, so your queries, documents, and extractions never leave your infrastructure.

Search across many sources from one screen and narrow fast with rich filters. Multiple sources, one query: medical knowledge bases plus PubMed (NCBI), Semantic Scholar, Europe PMC, and OpenAlex (250M+ scholarly works) - or your own uploaded documents (ZIP upload or an S3 prefix).

Keyword, semantic, or document-ID search modes, with MeSH-aware query expansion. Powerful filters: date range, journal, article type, language, author, open-access, and quality signals like impact factor, journal quartile, h-index, and citation counts.

How to use:

Requirement: This app requires a John Snow Labs medical LLM - either the DeepLens API (accessed with an API key) or a John Snow Labs medical LLM model deployed on your own Amazon SageMaker endpoint. Allow outbound internet access.

• Launch the AMI and configure the security group to allow inbound HTTP traffic on port 80. Make sure all outbound traffic is allowed.
• Wait approximately 3-5 minutes for the application to initialize, then open http://INSTANCE_PUBLIC_IP in your browser.
• Log in using the EC2 Instance ID as the password (available in the EC2 console). No username is required.
• On the Setup page:
• Select your provider (DeepLens or SageMaker).
• Enter your API key (if you selected DeepLens) or SageMaker endpoint respectivelly.
• Choose the model you want to use.
• Click Test Connection.
• Click Save.
• Start a literature review:
• Enter your research topic.
• Define the data points you want to extract.
• Click Run Review.
• Review the extracted results in the table.
• Export the results to CSV if needed.
• (Optional) Select a set of articles and use Literature Q&A to ask questions about the selected papers.
• (Optional) Under Manage API Keys, add external literature-source API keys for:
• PubMed
• Semantic Scholar
• Europe PMC
• OpenAlex

Note: If you need a DeepLens API key, please Contact John Snow Labs 

For additional information see Documentation