Launching today

SUPERWISE Sentinel

Launching today

The AI Gateway

26 followers

The AI Gateway

26 followers

Visit website

SUPERWISE Sentinel is an enterprise AI gateway. Mac, Windows, Linux – Deploy a localized proxy, capture 100% of your LLM traffic, and redact secrets, PII, and toxic content at the edge. Your raw data never leaves your perimeter.

Interactive

Free Options

Launch tags:Artificial Intelligence•Development•Security

Launch Team / Built With

Maker

📌

Hey Product Hunt, Pretty much every team we talk to has the same problem: AI got into the building way faster than anyone could write a policy for it. API keys in .env files, coding agents pasting stack traces (creds included) into three different providers, and security finding out about it after the invoice showed up. The standard fix is "add our SDK to every app." Which means touching every codebase, nagging every team, and still missing the Python script someone wrote on a Friday afternoon. So we built SUPERWISE Sentinel, an AI gateway that runs as a lightweight proxy at your network edge. Point your base URL at it and every LLM call (OpenAI, Anthropic, Gemini, or your own OpenAI-compatible model on vLLM) flows through it. No code changes. It's an env var swap. What it does today: Redacts secrets and credentials (API keys, cloud creds, tokens) from prompts before they leave. Redacts PII: emails, phone numbers, SSNs, card numbers, catches toxic and hostile content, logs every call to an Activity Log, so you have an actual audit trail, shows token usage by provider, model, and gateway (tokens, not dollars. We're not going to pretend we know your negotiated rate.) The design decision I care most about: split-plane. The proxy runs in your environment (laptop, Docker, cluster) and the control plane is SaaS. Your raw prompts never touch our servers, only anonymous metadata does. We wouldn't install a "security" tool that copies every prompt into someone else's cloud, and you shouldn't either. The install should only take 5 minutes. Give it a whirl and let us know what you think!

Report

20d ago

Maker

Hey Product Hunters,

I'm John, co-founder and CPO at SUPERWISE. We've spent years in AI governance, and the pattern that finally made us build Sentinel was simple: the data that leaks into AI tools isn't leaking from some rogue app. It's people doing their jobs. A support rep pastes a customer record, an engineer pastes a stack trace with keys in it.

Sentinel runs on your own machine or network, close to where your data and apps already live. Passwords, API keys, health data and customer details are redacted before anything reaches the model, and every decision is logged so you can show what was caught.

Windows has a signed one-click installer. Mac and Linux install in a few minutes from the deployment guide. Every plan is free for 30 days, no credit card. If you'd rather see it first, the live demo runs a prompt through it with no install: superwise.ai/sentinel/demo

One question for you: what's the thing your team is most worried about pasting into an AI tool? I'll answer every comment today.

Report

7h ago

Excited to see how Sentinel from Superwise can help healthcare enterprises protect their AI engagements for both business and quality of care applications.

Report

6h ago

Maker

Hey Hunters 👋
We're the team behind Superwise Sentinel — an AI Gateway that sits in front of your LLM traffic.

Point your apps at it (no code changes) and it redacts or blocks secrets, PII, and toxic content before they reach the model — with live visibility into everything. Your prompts and responses never leave your environment; only anonymous metadata reaches us.

Up and running in minutes via CLI, Docker, Helm, or a one-click Windows app. We'd love your feedback 🙏

Report

6h ago

Forum Threads

p/superwise-sentinel

•

20h ago

Launching Tomorrow!

Your engineers are already using AI all day. Claude Code in the terminal, LangChain inside the app, an OpenClaw session nobody mentioned in standup, and some script from March that still calls the OpenAI API directly.

None of it routes through anything and there's no front door. So three problems pile up at once and you can't fix any one of them on its own.

You can't see it. Nobody can answer "what did we send to an LLM last week, and who sent it" without a spreadsheet and a genuinely bad afternoon.

View all