Purfect Labs Learning Center

An interactive walkthrough of Shield's architecture: proxy flow, filter packs, threat model, and tamper-evident audit trail. Tab through each layer of the security stack.

Feature-by-feature comparison matrix: Shield vs. WAFs, prompt guardrails, and manual processes. Expandable rows, live code comparison, and alternative deep dives with pros/cons.

Live entropy calculator and regex comparison tool. Learn why entropy analysis catches custom tokens, encoded credentials, and novel secrets that regex patterns miss entirely.

Interactive threat scenarios and attack surface comparison. See how injection manipulates model behavior while exfiltration leaks sensitive data, and how Shield blocks both in one proxy.

Interactive guide mapping Shield's architecture to major compliance frameworks. See exactly which controls Shield satisfies with specific clause references and evidence mappings.

Interactive map of the AI supply chain. Click each node to see what leaks at every layer and toggle Shield on/off to see how local redaction closes every gap.

Interactive checklist covering the 10 most common LLM security gaps, from secret detection and compliance scope to incident response planning.

Interactive redaction sandbox, paste any text and toggle Shield's four filter packs to see exactly what gets caught. Every match comes with an explanation.

Interactive data flow simulator, toggle Shield on/off to see exactly what sensitive data reaches LLM training pipelines. Provider policy explorer included.

Interactive threat model explorer for Model Context Protocol security. Map attack vectors across tool impersonation, parameter injection, and credential exposure.

Interactive incident response guide for AI data breaches, explore three real-world scenarios, walk through a five-phase response timeline.

Interactive guide showing the architectural gap between network-layer WAF inspection and application-layer AI threats. Side-by-side threat coverage comparison.

Architectural comparison of self-hosted and API-based AI deployments. Threat model tabs, interactive decision matrix, and security gap analysis.

Interactive guide to AI red teaming, systematically test your LLM applications for prompt injection, data exfiltration, and supply chain vulnerabilities.

Paste any prompt and watch Shield's redaction engine catch PII, secrets, PHI, and payment data in real time. Toggle filter categories, see highlighted matches.

Interactive checklist covering 10 critical API key management practices, hardcoded keys, missing rotation, shared credentials, zero audit trail. Score your team's key hygiene.

Interactive walkthrough of the five Trust Services Criteria for AI systems. OWASP LLM Top 10 mapping, audit readiness checklist, and evidence collection guidance.

Three-scenario redaction sandbox for telecom: customer support with CPNI, network operations with device identifiers, and billing disputes with payment data. Compliance mapping for CPNI, CALEA, FCC 2024 breach rules, GDPR, NIST CSF, and ISO 27001:2022.

Learn how to securely authenticate and authorize AI agents. Covers OAuth2 service accounts, OWASP LLM08 Excessive Agency, NIST AI RMF governance, and practical patterns for agent-to-API access control.

Learn how to classify data for AI use, public, internal, confidential, and restricted. Interactive classification tool, data-type matrix, and self-assessment for enterprise teams sending data to ChatGPT, Claude, and other LLMs.

Traditional DLP tools catch structured data in databases and email, but they're blind to unstructured AI prompts containing PII, secrets, and confidential business data. Learn how AI-aware DLP closes the gap.

Your AI prompts cross borders you didn't know existed. Learn how data residency laws (GDPR, PIPL, DPDP Act, LGPD) affect where your AI data is stored and processed, and how to keep sensitive data within your jurisdiction.

Compare the four major AI governance frameworks, NIST AI RMF, ISO 42001, EU AI Act, and OWASP Top 10 for LLM, and assess which apply to your organization. Interactive framework comparison with applicability assessment.

Compare how OpenAI, Anthropic, and Google handle your prompts and data. What they log, what they train on, and how to lock it down, before your data leaves your machine.

An LLM firewall inspects, modifies, and blocks prompts and responses between your applications and AI models like ChatGPT, Claude, and Copilot. Learn how it differs from a WAF, what threats it stops, and how to deploy one.

Compare on-premises vs cloud AI deployment: data sovereignty, compliance, latency, cost, and control. See how a local security gateway gives you cloud convenience with on-prem data protection.

How data redaction actually works: the regex patterns, entropy detection, and machine learning approaches that catch PII, PHI, PCI, and secrets before they reach ChatGPT, Claude, or Copilot.

Stop API keys, tokens, database credentials, and customer secrets from reaching ChatGPT, Claude, Copilot, or any AI model. Shield keeps your secrets on your machine, SOC 2 and ISO 27001-ready for DevTools teams.

Protect VINs, GPS traces, firmware builds, calibration parameters, supplier contracts, and connected-vehicle customer data before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps automotive data on your machines, ISO/SAE 21434, UN R155, TISAX, GDPR, CCPA-ready.

Stop customer PII, payment card data, and order details from reaching ChatGPT, Claude, or any AI model. Shield keeps e-commerce customer data on your machines, PCI DSS, GDPR, CCPA-ready.

Protect student records, grades, IEP data, and directory information before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps student data on your machines, FERPA, COPPA, PPRA-ready.

Protect grid telemetry, SCADA configurations, control system data, and NERC CIP audit materials before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps your operational data on your machines.

Protect customer data, payment information, and proprietary models before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps financial data on your machines, PCI DSS, GLBA, SOX-ready.

Stop controlled unclassified information (CUI), ITAR technical data, and procurement-sensitive documents from reaching ChatGPT, Claude, Copilot, or any AI model. Shield keeps government data on your machines.

Protect patient data before it reaches ChatGPT, Claude, or any AI model. Shield stops PHI, medical records, and insurance IDs from ever leaving your hospital's network.

Stop policyholder PII, medical records, financial data, and claims information from reaching ChatGPT, Claude, or any AI model. Shield keeps insurance data on your machines, HIPAA, GLBA, NAIC AI Principles-ready.

Stop confidential case data, client names, and settlement details from reaching ChatGPT, Claude, or any AI model. Shield keeps attorney-client privileged information on your firm's machines.

Protect manufacturing trade secrets, CUI, ITAR-controlled technical data, and supplier contracts before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps proprietary industrial data on your machines.

Keep customer PII, API keys, and internal architecture docs out of AI prompts. Pass enterprise security questionnaires and protect your SOC 2, ISO 27001, and GDPR posture, without slowing down your team.

Compare the real cost of building an AI data security solution in-house versus deploying Shield. Interactive calculator covers engineering, infrastructure, compliance, and maintenance, estimate your total cost of ownership.

Interactive cost calculator, estimate the financial impact of an AI prompt leak based on your company size, industry, and the data types you send to LLMs.

Interactive risk calculator, estimate your organization's shadow AI exposure in 2 minutes. See your risk score, data leak probability, and estimated annual cost.

Interactive policy builder, answer 6 questions and get a complete, framework-mapped AI security policy. 8 sections, ready for your next security audit.

Self-assessment across 5 dimensions, governance, visibility, data protection, threat detection, and compliance. 10 questions, 3 minutes.

6 questions. 2 minutes. Score your AI pipeline's security posture and get a personalized Shield tier recommendation. No email required.

Interactive scorecard across six dimensions: time, engineering capacity, compliance, cost, maintenance, and strategic fit. Rate your organization and get a clear build/hybrid/buy recommendation.

Self-assessment checklist covering SOC 2, HIPAA, GDPR, ISO 27001, PCI DSS, and NIST AI RMF controls. Check off your AI compliance readiness and get an instant score across every framework.

Build an inventory of every AI tool your team uses. Categorize by type, data sensitivity, integration depth, and user scope, then get a weighted risk score and prioritized mitigation plan for each one.

Select your compliance frameworks, AI tool type, and data sensitivity to generate a tailored 40+ question security questionnaire for evaluating any AI vendor. Covers data handling, encryption, access controls, incident response, and compliance certifications.

Compare what OpenAI, Anthropic, Google, and Microsoft log from your prompts, training policies, retention windows, and compliance gaps. Learn how to keep sensitive data off provider servers with local-first redaction.

Architecture deep dive into four cryptographic building blocks for trustworthy AI: verifiable inference with ZK proofs, data provenance with Merkle trees, on-chain governance with smart contracts, and agent identity with W3C DIDs and Verifiable Credentials.

Protect clinical trial data, drug discovery IP, and regulatory documents before they reach any AI model. Interactive redaction scenarios for clinical operations, medicinal chemistry, and FDA submissions.

Protect scripts, talent contracts, and pre-release footage before they reach any AI model. Interactive redaction scenarios for script development, talent negotiations, and pre-release marketing.

Protect donor records, grant proposals, and client case files before they reach any AI model. Interactive redaction scenarios for grant writing, donor management, and client services.

Protect client tax returns, SSNs, bank account numbers, and confidential financial data before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps taxpayer data on your machines, IRS Pub 1075, GLBA, FTC Safeguards-ready.

Protect mortgage applications, SSNs, tenant screening data, and confidential transaction documents before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps client data on your machines, GLBA, FTC Safeguards, FCRA, RESPA-ready.

Protect shipment manifests, driver PII, cargo declarations, and supply chain partner data before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps logistics data on your machines, TSA, C-TPAT, FMCSA-ready.

Protect guest PII, credit card data, loyalty program records, and event contracts before they reach ChatGPT, Claude, Copilot, or any AI model. Shield keeps guest data on your machines, PCI DSS, GDPR, CCPA-ready.