Show IH: Sentry - AI Email Shield that saves me 69 hours/year (Pre-inbox filtering)

Hey Indie Hackers! 👋

Alexey here. Like many of you, I have a "public" email for my SaaS. And like many of you, my morning routine used to start with 20 minutes of deleting:

• "Quick question..." (Sales pitch)
• "Partnership opportunity..." (Link exchange)
• "URGENT INVOICE" (Phishing)

I calculated I was spending ~70 hours a year just triaging spam. That's two weeks of dev time lost. Existing filters (Gmail, Outlook) work after the email hits my inbox. I wanted something that stops it before.

So I built Sentry.

🚀 What is it? Sentry is an AI-powered proxy layer for your email. It sits in front of your real inbox (Gmail/Outlook) and filters incoming traffic using LLMs (Gemini 2.0) trained on B2B spam patterns.

🛠️ How it works (Technical):

1. MX Layer: Point your MX records to Sentry (or use our aliases).
2. AI Analysis: We analyze sender headers, SPF/DKIM, and body content.
3. Routing:

• Clean: Forwarded to your real inbox via SMTP.
• Spam: Rejected instantly.
• Suspicious: Quarantined for manual review (via a dashboard).

💰 Results so far:

• My inbox volume dropped by 80%.
• Phishing attempts (fake invoices/CEOs) are blocked before I see them.
• I saved ~1.5 hours last week alone.

🙏I need your feedback: We just launched the Beta. It's free for early users. I'm looking for founders who are drowning in cold outreach to stress-test our AI model.

Does the setup flow make sense?
Would you trust a proxy with your business email? (We don't store bodies, only metadata, but trust is key).

What features are missing?
Check it out: https://sentry.wr.io/

Let me know what you think! I'm hanging out in the comments all day. 👇