Take back control. Intercept, audit, and route your AI traffic.
Project description
🛡️ AI DevSec Gateway (formerly AI Network Blocker)
Zero-Trust interceptor, auditor, and router for all your AI traffic.
📖 What is this?
AI DevSec Gateway is an enterprise-grade, open-source proxy and DevSecOps tool that intercepts, audits, and routes AI traffic leaving your local machine.
Originally created as a simple GUI to block AI endpoints, it has evolved into a comprehensive Zero-Trust Gateway. It empowers developers and security teams to monitor exactly what data their AI coding assistants (like Copilot, Cursor, or extensions) are exfiltrating, intercept those requests, and route them to private, local, or corporate LLMs.
- Intercept & Block: A deterministic OS-level override via the
hostsfile that drops unauthorized outbound connections to 38+ AI domains. - Route: A transparent local HTTP proxy that intercepts cloud API requests and reroutes them to local LLMs (like Ollama, LM Studio, or vLLM).
- Audit: Real-time semantic analysis of active development environments to prevent data leakage and proprietary logic exposure.
✨ Features
| Feature | Description |
|---|---|
| 🔀 Transparent API Router | Seamlessly reroute Copilot/Cursor HTTP traffic to your own Local LLM inference servers. |
| 🛡️ AI DevSec Auditor | Live, socket-level analysis of running processes to detect telemetry leaks. Powered by on-demand OpenAI audits (Zero-Persistence). |
| 💻 Native CLI Interface | Full headless control for CI/CD environments. Use ai-blocker --status or ai-devsec-gateway --block. |
| 🔒 Deterministic Kill Switch | Hard OS-level blocking (127.0.0.1 redirection). No ambiguity, no reliance on DNS filtering servers. |
| 📦 Universal Distribution | Install via pip, brew, scoop, or as a portable single-file binary for Windows/Linux/macOS. |
| 🌍 Multilingual GUI | A premium Catppuccin Mocha interface with 10 supported languages and smart OS elevation (UAC/sudo). |
🎯 Supported Providers
The default interception engine targets 38+ domains across major providers:
| Provider | Key domains intercepted |
|---|---|
| 🟢 OpenAI | api.openai.com, chatgpt.com, platform.openai.com |
| 🟠 Anthropic | claude.ai, api.anthropic.com, anthropic.com |
| 🐙 GitHub Copilot | copilot.github.com, api.githubcopilot.com, telemetry.githubcopilot.com |
| 🔵 Google AI | gemini.google.com, aistudio.google.com |
| 🟦 Microsoft | copilot.microsoft.com, bing.com |
| 🔷 Meta AI | meta.ai, ai.meta.com |
| 🌊 Mistral / DeepSeek / xAI | mistral.ai, api.deepseek.com, api.x.ai |
The blocklist is dynamically configurable via
ai_blocker/constants.py.
🏗️ Architecture
AI DevSec Gateway operates at the boundary between your local development environment and the cloud.
graph TD
subgraph Local Environment ["Local Development Network"]
IDE[VS Code / Cursor / IDE]
CLI[ai-blocker CLI]
GUI[DevSec Gateway GUI]
Gateway((Local API Gateway <br> 127.0.0.1))
OS_Hosts[(OS Hosts File)]
end
subgraph External ["Target Inference"]
LocalLLM[Private Local LLM <br> Ollama / vLLM]
PublicCloud[Public Cloud <br> OpenAI / Anthropic]
end
IDE -->|1. Outbound API Request| OS_Hosts
OS_Hosts -->|2. Intercepted| Gateway
Gateway -->|3a. Route active| LocalLLM
Gateway -.->|3b. Audited pass-through| PublicCloud
Gateway -->|3c. Block active| Drop[Connection Refused]
For an in-depth dive into our modular structure, Deep Packet Inspection (DPI) plans, and Threat Models, read our Architecture Documentation.
🚀 Quick Start
1. Python Package (Pip)
The fastest way to get started with the headless CLI.
pip install ai-devsec-gateway
# Native CLI commands are now available globally:
ai-blocker --status
ai-devsec-gateway --block
ai-devsec-gateway --unblock
2. Package Managers (macOS & Windows)
macOS (Homebrew):
brew tap Akunimal/ai-devsec-gateway https://github.com/Akunimal/AI-Router-Blocker-AiO
brew install ai-devsec-gateway
sudo ai-blocker --status
Windows (Scoop):
scoop bucket add ai-devsec-gateway https://github.com/Akunimal/AI-Router-Blocker-AiO.git
scoop install ai-devsec-gateway
ai-blocker --status
3. Portable GUI Binaries
If you prefer a rich visual interface without installing Python:
- Visit the Releases page.
- Download the executable for your OS (
.exe, macOS binary, or Linux AppImage). - Run the application (it will automatically request Admin/sudo privileges when toggling the network switch).
🔒 Security Model
- Zero-Persistence BYOK: API keys for the semantic DevSec auditor are strictly kept in-memory. They are never written to disk, preventing supply-chain credential theft.
- Surgical OS Modifications: The engine uses targeted
sed-like parsing to inject# AI-Blockmarkers into the OS hosts file. It guarantees absolute isolation from your existing DNS mappings. - Isolated Telemetry: The application itself contains absolutely zero tracking, analytics, or hidden background phone-home mechanics.
🤝 Open Source & Governance
We believe that security tools must be 100% transparent. This project is built under strict open-source governance:
- Architecture Guide: Complete technical specifications.
- Contributing Guide: Standards and PR templates.
- Code of Conduct: We foster a welcoming community.
- Security Policy: Responsible vulnerability disclosure.
- License: MIT Licensed.
🗺️ Roadmap & Future Vision
We are actively evolving towards an enterprise Zero-Trust DLP Engine. Upcoming milestones include:
- Real-Time DLP Sanitization: On-the-fly regex and heuristics to strip PII before routing.
- eBPF Kernel Telemetry: Detecting
.git/configexfiltration directly at the Linux kernel level. - Confidential Computing: Running the Gateway within Trusted Execution Environments (TEEs) like Intel SGX.
Explore our ROADMAP.md to see the full vision.
Audit the unseen. Route the restricted. Trust no packets.
The DevSecOps Gateway for the AI era.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file ai_devsec_gateway-1.3.2.tar.gz.
File metadata
- Download URL: ai_devsec_gateway-1.3.2.tar.gz
- Upload date:
- Size: 33.5 kB
- Tags: Source
- Uploaded using Trusted Publishing? Yes
- Uploaded via: twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
d1ce3b19f57608b9d677e90848dbe4cd4d63033dc32eb5376a2148f17198a978
|
|
| MD5 |
ea7cb616a674db3d70708d85edaf2d32
|
|
| BLAKE2b-256 |
002cb0f97fad304ec001f9c561a55996e3b790a6306b3ee5dd512a54d4582255
|
Provenance
The following attestation bundles were made for ai_devsec_gateway-1.3.2.tar.gz:
Publisher:
publish.yml on Akunimal/AI-Router-Blocker-AiO
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
ai_devsec_gateway-1.3.2.tar.gz -
Subject digest:
d1ce3b19f57608b9d677e90848dbe4cd4d63033dc32eb5376a2148f17198a978 - Sigstore transparency entry: 1687308242
- Sigstore integration time:
-
Permalink:
Akunimal/AI-Router-Blocker-AiO@ccdb51c08342aeabf385f8bb8df1a622a410df70 -
Branch / Tag:
refs/tags/v1.3.2 - Owner: https://github.com/Akunimal
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
publish.yml@ccdb51c08342aeabf385f8bb8df1a622a410df70 -
Trigger Event:
release
-
Statement type:
File details
Details for the file ai_devsec_gateway-1.3.2-py3-none-any.whl.
File metadata
- Download URL: ai_devsec_gateway-1.3.2-py3-none-any.whl
- Upload date:
- Size: 29.0 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via: twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
6e4c97dea39de8ccdf3ccf69c51194b0c01b6beed33df4c921bfd0965a4cdae0
|
|
| MD5 |
6a09ea1cc991c9c36623cdc4499120f9
|
|
| BLAKE2b-256 |
58c9c04d688cdb543b8ffa9e0c3b42296419810b948491dd5cb79e6b7e01fef0
|
Provenance
The following attestation bundles were made for ai_devsec_gateway-1.3.2-py3-none-any.whl:
Publisher:
publish.yml on Akunimal/AI-Router-Blocker-AiO
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
ai_devsec_gateway-1.3.2-py3-none-any.whl -
Subject digest:
6e4c97dea39de8ccdf3ccf69c51194b0c01b6beed33df4c921bfd0965a4cdae0 - Sigstore transparency entry: 1687308310
- Sigstore integration time:
-
Permalink:
Akunimal/AI-Router-Blocker-AiO@ccdb51c08342aeabf385f8bb8df1a622a410df70 -
Branch / Tag:
refs/tags/v1.3.2 - Owner: https://github.com/Akunimal
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
publish.yml@ccdb51c08342aeabf385f8bb8df1a622a410df70 -
Trigger Event:
release
-
Statement type: