⚡ AegisGate Rampart v0.7.1 is LIVE — Local proxy, IDE plugins for VS Code/Cursor and JetBrains, real-time detection. Download from GitHub (free, open source, Apache 2.0).
Download Rampart v0.7.1 → Quick start guide →
What is Rampart?
Rampart is a local security tool for developers who use AI coding assistants like GitHub Copilot, Cursor, or local LLMs (Ollama, LM Studio). It sits between your editor and the AI model, scanning everything you send — and everything the AI sends back — for sensitive data and security risks.
Think of it as a firewall for AI coding tools. It catches the moment you’re about to send a database password to Copilot, or when the AI generates code that contains an API key, and stops it before it’s too late.
Do I need Rampart?
If you use AI chat in a browser (ChatGPT, Claude, etc.): You need Lens — it’s a free browser extension that protects you in the browser. Rampart is not for this use case.
If you use AI coding tools (Copilot, Cursor, local LLMs, API calls): You need Rampart. It protects the places Lens can’t reach — your IDE, your terminal, your API calls.
| Your setup | What to use |
|---|---|
| ChatGPT or Claude in a browser | Lens (free browser extension) |
| GitHub Copilot in VS Code | Rampart (IDE plugin) |
| Cursor AI editor | Rampart (IDE plugin) |
| Local LLMs (Ollama, LM Studio) | Rampart (local proxy) |
| API calls to OpenAI/Anthropic from your code | Rampart (local proxy) |
| Both browser chat AND coding tools | Lens + Rampart (both are free) |
See Rampart in Action
Watch Rampart intercept AI traffic from VS Code, detect hardcoded secrets, and protect your coding workflow.
What does it catch?
| Risk | Examples |
|---|---|
| Secrets | API keys (AWS, GitHub, OpenAI, Stripe), database passwords, SSH private keys, JWT tokens, OAuth tokens |
| Personal info (PII) | SSN, email, phone, credit card, passport, bank routing numbers |
| Prompt injection | Adversarial prompts designed to make the AI ignore safety rules, leak system prompts, or execute unauthorized actions |
| Compliance violations | Text that violates HIPAA, GDPR, PCI-DSS, EU AI Act |
| Malicious code (XSS) | Script injection, event handlers, encoded payloads, SVG vectors |
| Response risks | PII leaked in AI responses, hallucinated secrets, injected content in model output |
Architecture Overview
Detection Flow
Is it private?
Yes. Everything runs on your machine. Nothing is sent to any server — not your prompts, not your code, not your detection results.
| ✅ Rampart DOES | ❌ Rampart does NOT |
|---|---|
| Run entirely on your machine | Send your prompts or code to any server |
| Scan traffic locally in real-time | Phone home with analytics or telemetry |
| Encrypt its CA keys with a passphrase | Store plaintext keys on disk |
| Redact PII/secrets from audit logs | Log your actual prompt text or secrets |
| Stay free and open source | Require an account or subscription |
Open source (Apache 2.0): github.com/aegisgatesecurity/aegisgate-rampart
Quick Start
Option 1: Install the IDE plugin (recommended)
VS Code / Cursor:
- Open the Extensions panel (
Ctrl+Shift+X/Cmd+Shift+X) - Search for “AegisGate Rampart”
- Click Install, then Reload
- Detection runs automatically as you type — results appear as inline warnings
JetBrains (IntelliJ, PyCharm, WebStorm, etc.):
- Open Settings → Plugins → Marketplace
- Search for “AegisGate Rampart”
- Click Install, then Restart
- Detection results appear in the Problems tool window
Neovim / any LSP editor:
# Download the Rampart LSP binary
curl -L https://github.com/aegisgatesecurity/aegisgate-rampart/releases/latest/download/rampart-lsp-linux-amd64 -o /usr/local/bin/rampart-lsp
chmod +x /usr/local/bin/rampart-lsp
# Add to your LSP config (Neovim example)
# lspconfig.rampart_lsp.setup({})
Option 2: Run as a local proxy
# Download for your platform
curl -L https://github.com/aegisgatesecurity/aegisgate-rampart/releases/latest/download/aegisgate-rampart-linux-amd64 -o aegisgate-rampart
chmod +x aegisgate-rampart
# Start the proxy
./aegisgate-rampart --port=8443 --upstream=https://api.openai.com
# Point your AI tool at the proxy
# Instead of: https://api.openai.com/v1/chat/completions
# Use: http://localhost:8443/v1/chat/completions
Option 3: Docker
docker run -d \
-p 8443:8443 \
-p 9090:9090 \
ghcr.io/aegisgatesecurity/aegisgate-rampart:v0.7.1
How it works
Rampart runs in two modes — you can use either or both:
1. IDE Plugin Mode
Rampart runs inside your editor as a Language Server (LSP). As you type prompts or code that will be sent to an AI tool, Rampart checks for sensitive data and security risks in real-time. Warnings appear inline, just like lint errors.
| Editor | Plugin | Status |
|---|---|---|
| VS Code / Cursor | aegisgate-rampart-ext | v0.3.0, published (IDE plugins — Rampart core is v0.6.2) |
| JetBrains (IntelliJ, PyCharm, etc.) | aegisgate-rampart-jetbrains | v0.3.0, published (IDE plugins — Rampart core is v0.6.2) |
| Neovim / Emacs / Helix / Sublime | Rampart-LSP (generic LSP server) | Available via rampart-lsp binary |
The LSP server uses JSON-RPC 2.0 over stdio — it works with any editor that supports LSP.
2. Local Proxy Mode
Rampart runs as a transparent proxy on your machine. Any AI tool that makes HTTP requests — Copilot, Cursor, local LLMs, API calls — gets intercepted and scanned.
Your AI tool (Copilot/Cursor/API) → Rampart proxy (localhost:8443) → AI model (OpenAI/Anthropic/local)
- Rampart generates a local CA certificate and configures your system to trust it
- All HTTPS traffic to AI services flows through the proxy
- Requests and responses are scanned in real-time
- Malicious content is blocked before it reaches the model (or before the response reaches you)
- CA keys are encrypted at rest with a passphrase — no plaintext keys on disk
Rampart vs Lens vs Platform
| Feature | Lens (browser) | Rampart (local proxy/IDE) | Platform (server) |
|---|---|---|---|
| Protects browser AI chat | ✅ | — | — |
| Protects IDE AI tools | — | ✅ | — |
| Protects API calls | — | ✅ | ✅ |
| Protects server-to-server AI | — | — | ✅ |
| Real-time editor detection | — | ✅ | — |
| Compliance frameworks | 5 facets | 5 facets | 31 frameworks |
| SIEM/SOAR integration | — | — | ✅ |
| Multi-tenant | — | — | ✅ |
| Air-gapped deployment | — | — | ✅ |
| Price | Free | Free | Free tier + paid |
For Organizations
Typical team setups
- Individual developer: Lens (browser) + Rampart (IDE) → full local protection, zero cost
- Developer team (2-10): Rampart (IDE) for each dev + Platform (server) for centralized policy, audit logs, and compliance reporting
- Enterprise: Platform (server) + Lens (browser) for non-technical staff + Rampart (IDE) for developers + SIEM/SOAR integration + 31 compliance frameworks
Privacy & Security (for compliance teams)
- Zero data collection — no telemetry, no analytics, no phone-home
- All detection is local — patterns and ML model run on your machine
- CA keys encrypted at rest — passphrase-protected, never plaintext
- Audit log redaction — PII and secrets stripped from logs before writing to disk
- No network egress — Rampart never sends your data anywhere
- Open source — Apache 2.0, auditable, no hidden behavior
Download
Release Assets (v0.7.1)
| Platform | Asset | Architecture |
|---|---|---|
| macOS | aegisgate-rampart-darwin-amd64.tar.gz | Intel |
| macOS | aegisgate-rampart-darwin-arm64.tar.gz | Apple Silicon |
| Linux | aegisgate-rampart-linux-amd64.tar.gz | x86_64 |
| Linux | aegisgate-rampart-linux-arm64.tar.gz | ARM64 |
| Linux | aegisgate-rampart-linux-amd64.deb | Debian/Ubuntu |
| Linux | aegisgate-rampart-linux-amd64.rpm | RHEL/Fedora |
| Windows | aegisgate-rampart-windows-amd64.zip | x86_64 |
| Docker | ghcr.io/aegisgatesecurity/aegisgate-rampart:v0.7.1 | Multi-arch (cosign-signed) |
All 13 assets on GitHub → Source code →
IDE Integration Details
For detailed setup instructions for each editor, see our IDE Integration Guide.
VS Code / Cursor setup
- Install the AegisGate Rampart extension from the VS Code marketplace
- The extension automatically starts the Rampart LSP server
- Detection results appear as inline diagnostics (like lint warnings)
- Configure severity levels in
.vscode/settings.json:
{
"rampart.detection.severity": "warning",
"rampart.detection.categories": ["pii", "secrets", "xss", "compliance", "ml"]
}
JetBrains setup
- Open Settings → Plugins → Marketplace
- Search “AegisGate Rampart” → Install → Restart
- Detection results appear in the Problems tool window
- Configure in Settings → Tools → AegisGate Rampart
Neovim setup (any LSP editor)
-- Neovim (init.lua with nvim-lspconfig)
require'lspconfig'.rampart_lsp.setup{
cmd = {"/path/to/rampart-lsp"},
filetypes = {"*"}, -- scan all file types
settings = {
detection = {
categories = {"pii", "secrets", "xss", "compliance", "ml"},
severity = "warning"
}
}
}
Architecture
Copilot/Cursor/API] --> B[Rampart Proxy
localhost:8443] A --> C[IDE Plugin
VS Code/JetBrains/LSP] B --> D[Detection Engine] C --> D D --> E{Block or Allow?} E -->|Block| F[🛑 Warning + Log] E -->|Allow| G[Forward to AI Model] end G --> H[AI Model
OpenAI/Anthropic/Local] H --> I[AI Response] I --> B B --> J[Response Scanner] J --> K{Safe?} K -->|No| F K -->|Yes| L[Return to AI Tool]
Frequently Asked Questions
Does Rampart slow down my AI tools?
No. Detection adds ~5ms latency, which is imperceptible compared to the seconds-long response times of AI models. The ML model runs in pure Go — no external process, no network calls.
Does Rampart store my prompts?
No. Rampart processes prompts in memory and discards them. Audit logs are written with PII and secrets redacted — only the detection category and severity are recorded.
Can I use Rampart without the proxy?
Yes. The IDE plugin mode (LSP) works without the proxy. You can also use just the proxy without the IDE plugin. They’re independent.
How is Rampart different from Lens?
Lens is a browser extension that watches what you type into AI chat websites. Rampart is a system-level proxy and IDE plugin that catches AI traffic from any application — Copilot, Cursor, API calls, local LLMs. They use the same detection engine but protect different surfaces.
Is Rampart free for commercial use?
Yes. Apache 2.0 license. Free for personal and commercial use. No restrictions, no attribution required beyond the license terms.
Canonical facts (v0.7.1)
Source: aegisgate-rampart repo
- Local proxy: Intercepts AI traffic transparently at the system level — no app changes needed
- IDE integration: VS Code/Cursor extension, JetBrains plugin, generic LSP server (Neovim, Emacs, Helix, Sublime)
- MITM block mode: Blocks malicious prompts before they reach the AI model
- CA key encryption: Encrypted at rest with passphrase — no plaintext keys on disk
- Audit log redaction: PII and secrets stripped from logs before writing
- 1,318 test functions, 80.7% coverage
- 13 release assets: macOS (Intel + ARM), Linux (deb + rpm), Windows, Docker (multi-arch, signed)
- Apache 2.0, zero external dependencies for core functionality