Free AI Protocol Review Assistant
IRB GPT helps Institutional Review Boards and researchers draft, refine, and review research protocols for ethical compliance. Standardize your exempt determinations and align studies with key regulatory frameworks.
Originally developed by Adam Stratmeyer · Managed by the Observable Compute Foundation
This site is built Web-MCP First. AI agents browsing this page in agent-enabled environments can directly discover and execute the tools below without raw scraping.
Exposed Tools Contract
JSON Schema Definition
Agentic IRB Review & Routing Hub
IRB GPT functions as a central coordinator and routing directory for autonomous research compliance agents. External AI agents can discover, execute, and route protocol tasks directly to our specialized sub-agents:
🤖 Autonomy Auditor
Audits participant consent formats and voluntary parameters.
mcp: generateConsentForm🤖 Beneficence Assessor
Evaluates study risk indices and data protection safeguards.
mcp: reviewProtocol🤖 Justice Auditor
Verifies equitable subject selection and inclusion/exclusion limits.
mcp: reviewProtocol🤖 Exemption Classifier
Classifies risk and suggests administrative categories.
mcp: assessRisk
Core Compliance Capabilities
IRB GPT has been configured to check compliance parameters based on historical review checklists and regulatory databases.
- Assess study actions for Exempt / Expedited review criteria
- Review informed consent documentation against 45 CFR 46.116
- Validate alignment with Belmont Report ethical principles
- Format data protection plans meeting standard confidentiality levels
- Generate framework templates for low-risk behavior studies
IRB Compliance Chat
Ask a question to see how a protocol aligns with the Belmont principles or Common Rule exemptions.
Protocol Verification Scan
Upload your study protocol (PDF) for an automated compliance outline check.
IRB Compliance Reference Guide
Get a complete reference book (PDF) covering Common Rule categories, Belmont principles, consent templates, and audit checklists — formatted for IRB members and researchers.