Skip to content
TruthVouch Docs

TruthVouch Platform

Welcome to the TruthVouch Platform. Govern your AI systems with real-time scanning, policy enforcement, and content verification.

Overview

TruthVouch provides a unified platform for AI governance with six core modules:

1. AI Firewall (Governance Gateway)

Scan and protect all AI requests and responses in real-time.

Key features:

  • PII, injection, toxicity detection
  • Policy enforcement
  • Compliance checking (GDPR, HIPAA, SOC2)
  • Hallucination detection
  • Multi-provider support (OpenAI, Anthropic, Azure, etc.)

Deploy Firewall →

2. Sentinel Agent (Endpoint Monitoring)

Monitor and enforce AI tool usage on employee devices.

Key features:

  • Desktop monitoring (ChatGPT, Copilot, Gemini, Claude)
  • Local DLP policy enforcement
  • Offline-first architecture
  • Encrypted telemetry reporting
  • Cross-platform (Windows, macOS, Linux)

Deploy Sentinel →

3. Governance Portal (Dashboard)

Centralized control for policies, alerts, and analytics.

Key features:

  • Policy management
  • Real-time alerts and investigation
  • Usage analytics
  • Compliance reporting
  • Team collaboration

Open Dashboard →

4. Truth Knowledge Base (Nuggets)

Certify and manage authoritative content.

Key features:

  • Create verified fact nuggets
  • Cryptographic certification
  • Blockchain transparency (optional)
  • Hallucination detection against nuggets
  • Multi-source support

Manage Nuggets →

5. Compliance & Audit (Trails)

Track all governance decisions for compliance and investigation.

Key features:

  • Complete audit trails (90+ days)
  • Compliance reports (SOC2, GDPR, HIPAA)
  • Decision investigation
  • Historical analysis
  • Export and integration

View Audit Trail →

6. API & Webhooks (Integration)

Build custom integrations on top of TruthVouch.

Key features:

  • RESTful Platform API
  • Webhook events
  • gRPC scanning interface
  • Custom policy rules
  • SDKs and libraries

Explore APIs →

Platform Architecture

┌─────────────────────────────────────────────────────────┐
│            TruthVouch Dashboard (SaaS)                   │
│  • Policy Management  • Analytics  • Alerts & Reporting │
└────────────────┬────────────────────────┬───────────────┘
                 │                        │
         Policy Sync (hourly)        Webhook Events
                 │                        │
    ┌────────────▼──────────┐      ┌─────▼──────────────┐
    │  Your Infrastructure  │      │  Your Systems      │
    ├──────────────────────┤      ├────────────────────┤
    │ Governance Gateway   │      │ Sentinel Agent     │
    │ • Scan LLM traffic   │      │ • Monitor AI tools │
    │ • Enforce policies   │      │ • Local DLP        │
    │ • Real-time alerts   │      │ • Encrypted sync   │
    │                      │      │                    │
    │ PostgreSQL + pgvector│      │ SQLite cache       │
    └──────────┬───────────┘      └────────────────────┘
               │
      ┌────────▼──────────┐
      │   LLM Providers   │
      │ • OpenAI          │
      │ • Anthropic       │
      │ • Azure OpenAI    │
      │ • Local models    │
      └───────────────────┘

Getting Started

1. For SaaS Users

Start protecting your AI systems immediately:

  1. Sign up for TruthVouch
  2. Create organization and invite team members
  3. Deploy Governance Gateway (Docker/Kubernetes)
  4. Integrate with your LLM APIs
  5. Configure policies in dashboard

SaaS Setup Guide →

2. For Self-Hosted Deployments

Deploy Governance Gateway and Sentinel on your infrastructure:

  1. Deploy Gateway with Docker
  2. Deploy Gateway with Kubernetes
  3. Install Sentinel on Windows/macOS/Linux
  4. Configure policies locally
  5. Monitor with Prometheus/Grafana

Self-Hosted Guide →

Core Features

Real-Time Scanning

Every AI request and response is scanned for:

  • PII — SSN, credit cards, email addresses
  • Injection attacks — Prompt injection, SQL injection
  • Toxicity — Harmful, hateful, or inappropriate content
  • Hallucinations — False or unsupported claims
  • Compliance violations — GDPR, HIPAA, SOC2 requirements

Learn about scanning →

Policy Enforcement

Define and enforce organizational AI governance:

  • AI tool allowlists — Which tools employees can use
  • Data protection — Block sensitive data from being shared
  • Time-based access — Restrict AI use during certain hours
  • Risk scoring — Flag high-risk requests
  • Custom rules — Build policies for your industry

Create policies →

Hallucination Detection

Identify when AI models generate false claims:

  • Semantic matching — Compare responses against verified sources
  • Contradiction detection — Find claims that contradict known facts
  • Confidence scoring — Rate likelihood of hallucination
  • Knowledge integration — Use your Truth Nuggets for verification

Understand hallucinations →

Compliance & Audit

Meet regulatory requirements with complete audit trails:

  • GDPR compliance — Track data processing and user consent
  • HIPAA compliance — Protect health information
  • SOC2 compliance — Document AI governance controls
  • Custom reports — Generate compliance documentation
  • Investigation — Trace any AI governance decision

Generate compliance reports →

Common Use Cases

Protect Customer Data

Prevent PII from being sent to public LLM APIs:

[Customer Data] → Gateway scans → [Allows if no PII] → OpenAI
                                → [Blocks if PII found]

Enforce Company Policies

Block unapproved AI tools:

Employee uses Gemini → Sentinel detects → Blocks per policy

Verify Content Accuracy

Detect hallucinations in AI responses:

ChatGPT response → Gateway compares → Truth Nuggets
                → Marks if contradictions found

Maintain Compliance

Document all AI governance decisions:

Every scan → Audit log → Export for compliance review

Team Collaboration

Organize by Role

RoleCapabilities
OwnerFull access, manage billing, delete organization
AdminManage members, policies, configure gateway
EditorCreate/edit policies, manage alerts
MemberView dashboards, acknowledge alerts
ViewerRead-only access

Share Dashboards

Create team dashboards with custom visualizations:

  • Real-time alerts
  • Usage by department
  • Compliance status
  • Top threats

Integration & Automation

Webhooks

Receive real-time events when threats are detected:

# PII detected in LLM response
{
  "event_type": "alert.detected",
  "detection_type": "pii",
  "pii_type": "ssn",
  "confidence": 0.95,
  "action_taken": "blocked"
}

Webhook documentation →

APIs

Build custom integrations:

from truthvouch import Gateway


gateway = Gateway(api_key="...")
response = gateway.scan(
    request_body="user prompt",
    model="gpt-4"
)
print(f"Safe: {response.allowed}")

API reference →

SDKs

Integrate directly in your code:

  • Python SDK
  • Node.js SDK
  • .NET SDK
  • Go SDK

Performance & Scalability

Latency

  • Scan latency: 10-100ms (typical)
  • Gateway overhead: 50-200ms per request
  • Multi-region: Deploy in multiple zones

Throughput

  • Single instance: 500+ requests/second
  • Horizontal scaling: Add instances linearly
  • Auto-scaling: Kubernetes HPA support

Reliability

  • 99.99% uptime SLA
  • Automatic failover
  • Data replication
  • Offline operation (Sentinel)

Support & Documentation

Getting Help

  • Documentation: This site
  • Dashboard Help: In-app chat support
  • Email: support@truthvouch.io
  • Slack: For enterprise customers

Community

  • GitHub: truthvouch/examples
  • Discussions: GitHub Discussions
  • Blog: TruthVouch news and updates

Next Steps

Choose your path forward:

Just starting? → Take the interactive tour

Ready to deploy? → Self-Hosted Installation Guide

Want to integrate? → API & Webhooks Guide

Need compliance? → Compliance & Audit Trail