Works with Claude Code and Claude Desktop

Compliance verdicts, inside Claude

Give Claude access to your audit trail, compliance status, and PHI reports. Code and check compliance in the same session. No context-switching.

Get Early Access See it in action

Local-first — your data never leaves the machine

Permission-controlled — you approve every tool call

Works in Claude Code — ask while you code

Two products, one workflow

Claude Code captures what happens. Verdict evaluates it. Ask Claude about compliance — right in your coding session.

Capture

Claude Code writes code. The Verdict hook captures every session — prompts, file changes, commands, git context — into structured Change Records.

Evaluate

Verdict runs PHI detection, policy evaluation, and compliance mapping locally. Every session gets a pass/fail verdict against SOC 2, HIPAA, and HITRUST controls.

Query (same session)

Ask Claude Code about compliance mid-session. It calls Verdict tools to get real-time answers from your local data. No context-switching.

CaptureEvaluateQuery

See it in action

Mid-session compliance check in Claude Code. Describe what you need, Claude takes action.

Claude Code — Verdict MCP Active

You

I'm about to merge PR #42 for the patient intake form. Is it compliant?

Let me check the compliance status for PR #42.

verdict.get_compliance_verdict

PASS

Sessions

2 captured

Developer

jane@healthtech.com

Agent

Claude Code v2.1

Ticket

EHR-1847

PHI Risk

None

Duration

45 min

SOC 2 CC8.1 — Change authorized via EHR-1847

HIPAA §164.312(b) — Audit trail complete

HITRUST 09.b — Documentation present

PR #42 passes all compliance checks. Two sessions were captured totaling 45 minutes. The change is authorized via ticket EHR-1847, no PHI was detected, and all framework controls are satisfied. Safe to merge.

You

Were there any PHI issues across the team this week?

verdict.check_phi_exposure

Period

Mar 14–20, 2026

Sessions

47 scanned

Detections

Risk Level

None

Clean week. 47 sessions across the team with zero PHI detections. No action needed.

Describe the outcome. Claude takes action.

Six tools that Claude can call on your behalf. You approve each one. Results come from your local Verdict database in real time.

get_compliance_verdict

Compliance Verdict

Full pass/fail status for a PR or branch — policy results, control mappings, and evidence summary.

Try saying

“Is PR #42 ready to merge?”

check_phi_exposure

PHI Exposure Check

Scan sessions for PHI in prompts, file reads, or modifications. Risk levels and redaction status.

Try saying

“Any PHI risks this sprint?”

get_audit_trail

Audit Trail

Complete audit log for a session, developer, or time range. Formatted for auditor review.

Try saying

“Show the audit trail for last week”

get_change_record

Change Record

Full provenance for a session — developer, agent, model, files modified, commands run, git context.

Try saying

“What happened in session abc123?”

query_sessions

Session Query

Search and filter sessions by any dimension — date, developer, repo, status, or PHI detections.

Try saying

“How many sessions this month?”

export_compliance_report

Compliance Report

Audit-ready report mapped to SOC 2, HIPAA, or HITRUST controls. JSON or summary format.

Try saying

“Generate a SOC 2 report for Q1”

Set up in 2 minutes

Install Verdict, add one config block, restart Claude. Done.

Install Verdict

Terminal

$ brew install verdict

Add to Claude Code

.claude/settings.json

{
  "mcpServers": {
    "verdict": {
      "command": "verdict",
      "args": ["mcp", "serve"]
    }
  }
}

Start asking

Start a Claude Code session. Verdict tools are available immediately. Ask about compliance anytime mid-session.

Built for every role

Different people, same interface. Ask Claude what you need in plain English.

Developer

Check compliance before merging

"Is my PR compliant? Can I merge?" — get a pass/fail in seconds without opening a dashboard.

Tech Lead

Monitor team compliance posture

"Show me all sessions with policy warnings this sprint" — spot patterns and address issues early.

Compliance Officer

Query audit evidence on demand

"Generate a HIPAA compliance report for March" — structured evidence, no developer interruptions.

Auditor

Review evidence mapped to controls

"Show me SOC 2 CC8.1 evidence for Q1" — get exactly the documentation needed, instantly.

Model Context Protocol

MCP is Anthropic's open standard for connecting AI to external tools. Think of it as a USB-C port for AI — one universal interface that works across Claude Code, Claude Desktop, and any MCP-compatible client.

Verdict publishes an MCP server. Claude knows how to call it. You get compliance answers without leaving your conversation.

Build once, work everywhere

Same MCP server powers Claude Desktop, Claude.ai, and CLI integrations.

Local-first by design

All data stays on your machine. PHI never crosses the network.

Permission-controlled

Claude asks before every tool call. You approve or deny each action.

Real-time data

Always reads from your live Verdict database — not stale exports.

Get Verdict for Claude

Join healthcare IT teams using Verdict in Claude Code to stay compliant while shipping faster with AI agents.

Request Early Access View Full Product