GitHub - hoophq/hoop: The only access proxy that blocks dangerous linux commands and scrubs sensitive database outputs

hoop.dev

🔒 Access any database or server. Customer data automatically hidden. Everything recorded.

The only access proxy that blocks dangerous linux commands and scrubs sensitive database outputs

Without Hoop

-- Debugging production issue...
SELECT * FROM users WHERE id = 42;
┌────┬─────────────────────────┬─────────────┬─────────────────┬────────────────────┐
│ id │ email                   │ ssn         │ phone           │ credit_card        │
├────┼─────────────────────────┼─────────────┼─────────────────┼────────────────────┤
│ 42 │ john.doe@example.com    │ 123-45-6789 │ (555) 123-4567  │ 4111-1111-1111-1111 │
└────┴─────────────────────────┴─────────────┴─────────────────┴────────────────────┘
-- You screenshot the result for Slack...
-- 💀 SSNs, credit cards, and phone numbers now in your team chat

One query, one screenshot, one data breach.

With Hoop

Same query through Hoop:

-- You see:
┌────┬──────────────────┬─────────────┬──────────────────┬─────────────────┐
│ id │ email            │ ssn         │ phone            │ credit_card     │
├────┼──────────────────┼─────────────┼──────────────────┼─────────────────┤
│ 42 │ **************** │ *********** │ ************     │ *************** │
└────┴──────────────────┴─────────────┴──────────────────┴─────────────────┘

-- Now it's safe to share

Without Hoop

-- Fixing bug at 3AM...
UPDATE users SET name = 'Bob Ross'

-- 💀 1000000 rows updated

With Hoop

-- Fixing bug at 3AM...
UPDATE users SET name = 'Bob Ross'

-- 🚫 Query blocked by Guardrail: "Prevent UPDATE without WHERE

That's it. Hoop sits between you and your infrastructure. Sensitive data gets masked automatically. Dangerous operations blocked. Everything gets recorded.

30-Second Demo

# create a jwt secret for auth
echo "JWT_SECRET_KEY=$(openssl rand -hex 32)" >> .env

# download and run
curl -sL https://hoop.dev/docker-compose.yml > docker-compose.yml
docker compose up

View full installation options

How It Works

You → Hoop → Your Infrastructure
       ↓
   • Masks sensitive data (ML-powered)
   • Blocks dangerous commands
   • Records everything (for compliance)  
   • Controls access (who, what, when)

Works with:

• Databases: PostgreSQL, MySQL, MongoDB, Redis
• Servers: SSH, Kubernetes, Docker
• Tools: HTTP APIs, internal services

Why Teams Love Hoop

🧠 Smart Masking

Not regex. Machine learning that understands context.

• Knows "555-1234" is a phone number in user data
• Knows "BUILD-555-1234" is a build number
• Works in any language

⚡ Actually Fast

• <5ms latency
• No performance impact
• Works with existing tools

🔐 Real Security

• Nothing to configure
• Full audit trail
• SOC2/HIPAA/GDPR compliant

📚 Popular Guides

Databases

• MySQL
• PostgreSQL
• MongoDB
• MSSQL

Cloud & Infrastructure

• Kubernetes
• AWS
• SSH Jump Hosts

View all guides

Installation

Docker

bash

# create a jwt secret for auth
echo "JWT_SECRET_KEY=$(openssl rand -hex 32)" >> .env

# download and run
curl  -sL https://hoop.dev/docker-compose.yml > docker-compose.yml &&  docker compose up

See Docker Compose installation documentation

Kubernetes

See Kubernetes Deployment Documentation

AWS

See AWS Deploy & Host Documentation

Region	Launch Stack
N. Virginia (us-east-1)
Ohio (us-east-2)
N. California (us-west-1)
Oregon (us-west-2)
Ireland (eu-west-1)
London (eu-west-2)
Frankfurt (eu-central-1)
Sydney (ap-southeast-2)

View all regions

Advanced Features

What makes Hoop unique is its ability to not only inspect but also modify connections between users and infrastructure:

• AI Data Masking - Automatically hide sensitive data like emails, SSNs, and credit cards
• Just-in-Time Reviews - Approve risky commands in real-time through Slack or MS Teams
• Runbooks - Create pre-approved workflows for common tasks
• Web & Native Modes - Use the web interface or connect through your native database tools

See all features

You'll be in Good Company

• 200+ successful deployments from companies around the world
• 4.3 minute average setup time across all deployments
• Trusted by teams from startups to enterprises

🤝 Contributing

We welcome contributions! Check out our Development Documentation to get started.

📣 Community

Join our Discussions to ask questions, share ideas, and connect with other users.

⭐ Star this if you've ever worried about screenshots in Slack

Backed by