🛡️ AegisGate — Secure Every AI Interaction

Platform Version Lens Version Rampart Version License

Every time you type a prompt into an AI tool, you might be leaking sensitive data — passwords, API keys, personal information, trade secrets. AegisGate catches those risks before you press send. Three products, one mission: secure every AI interaction.

Open-core. Apache 2.0 community + commercial enterprise modules. No data collection. No vendor lock-in. Self-hosted only. Deploy in 60 seconds.

Complete AI Security Suite

From individual users to enterprise teams — comprehensive protection at every layer.

flowchart LR subgraph Individual["🛡️ Individual Users"] A[Browser Extension] A1[Lens] A2[10 AI Chat Tools] A3[Zero Data Collection] end subgraph Developer["⚡ Developers"] B[Local Proxy + IDE] B1[Rampart] B2[Copilot/Cursor/LLMs] B3[Real-time Detection] end subgraph Enterprise["🏢 Enterprise"] C[Security Gateway] C1[Platform] C2[6 Attack Surfaces] C3[31 Frameworks] end subgraph Core["🔐 Core Detection Engine"] D[176+ Patterns] D1[PII + Secrets] D2[XSS + Compliance] D3[ML Adversarial] end Individual --> Core Developer --> Core Enterprise --> Core style Individual fill:#1a1f2e,stroke:#38bdf8,stroke-width:2px style Developer fill:#1a1f2e,stroke:#f59e0b,stroke-width:2px style Enterprise fill:#1a1f2e,stroke:#10b981,stroke-width:2px style Core fill:#1a1f2e,stroke:#8b5cf6,stroke-width:2px

Three Products, One Mission

Pick your path — we'll meet you where you are.

🛡️ Lens — For Everyone FREE

You use ChatGPT, Claude, or other AI chat tools.

Lens is a free browser extension that watches what you type into AI chat sites. If you're about to send your Social Security number, a password, an API key, or a prompt injection attack — it warns you before you click send.

  • Works with 10 AI chat tools (ChatGPT, Claude, Gemini, Copilot, and more)
  • Detects PII, passwords, API keys, XSS, compliance risks, and adversarial attacks
  • 100% on-device — nothing ever leaves your browser
  • Free forever. No account needed. No data collected.

Install for Chrome → Install for Firefox → Learn more →

⚡ Rampart — For Developers FREE

You write code and use AI coding tools like Copilot or Cursor.

Rampart runs as a local proxy on your machine, intercepting AI traffic from any tool — Copilot, Cursor, local LLMs, API calls. It scans prompts and responses for security risks in real-time, right in your IDE.

  • Local proxy mode: intercepts all AI traffic transparently
  • IDE integration: VS Code, Cursor, JetBrains, Neovim, any LSP editor
  • Real-time detection in your editor — no context switching
  • Catches secrets, PII, injection attacks before they reach the AI model
  • Free and open source (Apache 2.0). No account needed. No data collected.

Download Rampart → Learn more →

🏢 Platform — For Teams & Organizations

You need to secure AI usage across your entire organization.

Platform is an enterprise gateway that sits between your users and any AI service. It scans all AI traffic — HTTP APIs, MCP, agent-to-agent protocols, and AI responses — with 31 compliance frameworks, SIEM/SOAR integration, and a web UI.

  • 6 attack surfaces: HTTP API, MCP, A2A, ACP, Response, Trust Framework
  • 31 compliance frameworks (HIPAA, SOC 2, EU AI Act, FedRAMP, and more)
  • SIEM integration (11 platforms), SOAR integration (4 platforms)
  • Web UI, multi-tenant, air-gapped deployment, single 19.1MB Go binary

🚀 Try the Live Demo Learn more →

What does AegisGate catch?

The risks you didn't know you were taking.

🔐 Personal Information (PII)

Social Security numbers, email addresses, phone numbers, passport numbers, credit card numbers, bank routing numbers. If it identifies a person, Lens and Rampart catch it before it reaches the AI.

🔑 Secrets & Credentials

API keys, AWS access keys, GitHub tokens, database passwords, private SSH keys, JWT tokens. The #1 way data breaches happen is secrets accidentally pasted into AI chat.

💉 Prompt Injection Attacks

Adversarial prompts designed to manipulate AI into ignoring safety rules, leaking system prompts, or executing unauthorized actions. Our ML model catches 100/100 known attack patterns.

📋 Compliance Risks

Text that violates HIPAA, GDPR, PCI-DSS, EU AI Act, or 27 other regulatory frameworks. Enterprise teams get automatic compliance scanning on every AI interaction.

Enterprise Differentiators: Verifiable AI Security

For organizations that need cryptographic proof, not just compliance checklists.

🔐 Trust Framework

Per-agent cryptographic identity with ECDSA P-256.

Every agent gets a unique keypair. Every action is signed. Every attestation is independently verifiable — no server round-trip needed.

  • Capability contracts enforced at runtime (fail-closed)
  • Real-time trust scoring (0-100) with anomaly detection
  • Signed attestations verifiable with OpenSSL
  • FIPS 140-2 crypto, air-gap compatible

Learn more →

🛡️ Federated IOC Store

Crowdsourced AI threat intelligence via signed bundles.

When one customer detects a novel attack, every customer is protected within minutes. Network effects make the platform stronger with every detection.

  • AI-specific IOCs (prompt patterns, adversarial examples)
  • Real-time distribution (< 5 minutes)
  • Cryptographically signed bundles (ECDSA P-256)
  • TAXII 2.1 + STIX 2.1 compatible

Learn more →

📋 AIBOM Generator

AI Bill of Materials for EU AI Act compliance.

Automatically generate CycloneDX 1.6+ SBOMs extended with AI components. Track models, prompts, training data, and embeddings with cryptographic integrity.

  • CycloneDX 1.6+ compliant (emerging standard)
  • Tracks models, prompts, datasets, embeddings
  • Byte-stable, cryptographically signed
  • PDF export for regulators, JSON for tooling

Learn more →

📊 CISO Posture Digest

Executive-ready PDF reports with multi-framework mapping.

Automated, signed reports for CISOs, boards, and auditors. HIPAA, PCI-DSS, SOX, EU AI Act compliance in plain language — no dashboard login required.

  • Executive summary with posture score (0-100)
  • Multi-framework compliance mapping
  • Trend analysis (6-month rolling)
  • Cryptographically signed, verifiable offline

Learn more →

Three products, one mission

🛡️ Lens⚡ Rampart🏢 Platform
Who it's forEveryoneDevelopersTeams & Enterprises
What it protectsBrowser AI chatIDE + local proxyAll AI traffic
DeploymentBrowser extensionLocal proxy + IDE pluginSelf-hosted server
AI tools covered10 chat toolsCopilot, Cursor, any LLMAny HTTP/MCP/A2A traffic
PriceFree foreverFree / open sourceFree tier + paid tiers
Compliance frameworks5 facets (PII, secrets, XSS, compliance, ML)Same detection engine31 frameworks
Data collectionZeroZeroZero (self-hosted)
Open source✅ Apache 2.0✅ Apache 2.0✅ Apache 2.0

Why AegisGate?

🔒 Zero Data Collection

Every product runs on your machine or your infrastructure. No telemetry. No analytics. No phone-home. Your prompts never leave your control.

🧠 ML Threat Detection

CharCNN-BiLSTM neural network (1.58M params) runs in pure JavaScript in your browser. 100/100 evasion resistance, 0% false positive rate, ~5ms inference. No cloud calls.

🏗️ Self-Hosted, No Lock-In

Platform is a single 19.1MB Go binary. Deploy in 60 seconds with Docker. Air-gapped deployment supported. No SaaS dependency, no vendor lock-in, no subscription required.

📊 Proven at Scale

5,000 concurrent virtual users, 23,578 requests per second, 0% errors. 8,000+ tests passing. 31 compliance frameworks. 11 SIEM platforms. 4 SOAR integrations.

Verifiable Security vs. Competitors

The four pillars no competitor can replicate.

CapabilityAegisGateCompetitor ACompetitor BCompetitor C
Per-agent cryptographic identity✅ ECDSA P-256❌ Anonymous❌ API key only⚠️ Shared service account
Crowdsourced threat intelligence✅ Network effect❌ Isolated❌ Vendor-curated only❌ None
AI bill of materials✅ CycloneDX 1.6+❌ Software SBOM only❌ No❌ No
Executive compliance reports✅ Signed PDF❌ Dashboard only⚠️ CSV export❌ No
Offline verification✅ Yes (OpenSSL)❌ Requires login❌ Requires login❌ N/A
Air-gap deployment✅ Full support❌ Cloud-only❌ Cloud-only⚠️ Limited
Open source✅ Apache 2.0❌ Proprietary❌ Proprietary❌ Proprietary

By the numbers

10

AI chat tools Lens protects

100/100

Adversarial attack patterns caught

0%

False positive rate

31

Compliance frameworks mapped

23,578

Requests per second sustained

~5ms

Detection latency

Get started in 60 seconds

🛡️ Install Lens

Free browser extension. Chrome or Firefox.

Chrome Web Store → Firefox Add-ons →

⚡ Download Rampart

Free local proxy + IDE plugin. macOS, Linux, Windows.

GitHub Releases →

VS Code, JetBrains, Neovim, any LSP editor

🏢 Deploy Platform

Single binary. Docker or bare metal.

docker run -d -p 8080:8080 \
  ghcr.io/aegisgatesecurity/aegisgate-platform:v4.4.0

Try Live Demo →

Explore further

📖 Documentation

🔗 Product Pages

  • Lens — browser extension for AI chat protection
  • Rampart — local proxy + IDE plugin for developers
  • Platform — enterprise gateway with verifiable security
  • Pricing — 5 tiers from free to enterprise
  • Competitive comparison — 7 competitors, 31 frameworks

🛡️ Security & Compliance

💬 Community