Back to browse
GitHub Repository

Local Privacy Firewall for AI

9 starsGo

Velar – Local privacy firewall for AI

by ubcent·Feb 25, 2026·2 points·0 comments

AI Analysis

●●SolidSolve My ProblemDark Horse

Regex-based PII masking for ChatGPT leaks, but detection is intentionally basic.

Strengths
  • Streaming-safe MITM proxy with deterministic placeholder restoration keeps apps working seamlessly
  • 100% local processing means zero data leaves machine for analysis or logging
  • macOS notifications and system proxy integration make adoption friction-low
Weaknesses
  • Regex-only detection is inherently limited—misses context-aware leaks and has high false-positive risk
  • No evidence of effectiveness against real-world data exfiltration patterns or testing against competitors
Category
Target Audience

Developers and teams using AI tools with access to sensitive data

Similar To

Nightfall AI · Pangolin (identity firewall) · Twingate

Post Description

I realized I was leaking sensitive data to ChatGPT every day. Emails, API keys, internal data. So I built a local firewall that detects and masks sensitive data before it leaves your machine. It runs as a local proxy, supports streaming, and restores data in responses. Everything is local. Would love feedback.

Similar Projects