Your data stays yours.
Slipstream only reads your database and only sees what your alerts ask for. With the Slipstream Agent, nothing connects in, so there’s no opening in your firewall; for a cloud database you can instead allow just Slipstream’s one fixed address. This page is written for you and for your IT department.
Slipstream can’t change your data
A read-only login
You give Slipstream a database login that can only read, such as membership of db_datareader on SQL Server.
Only SELECT queries
Queries are generated from what you build with clicks, never typed by hand, and anything other than a single SELECT is refused.
Read-only mode, every time
Each query runs in a read-only mode: on PostgreSQL the database itself refuses any change, and on SQL Server Slipstream undoes anything a query might change before it finishes.
For databases behind your firewall: nothing connects in
Many business databases are closed off from the internet: on a server in the office, or in the cloud inside a private network or behind a firewall. That’s good.
The Slipstream Agent is a small program you run on a computer in that network, an office server or a cloud virtual machine. It makes outgoing connections to Slipstream only, the same kind a web browser makes. Nothing connects in, so there’s no port to open in your firewall.
- The agent asks Slipstream for work, over outgoing https.
- Slipstream replies with the query to run, on that same connection.
- The agent runs it, read-only, against your database.
- The agent sends the rows back.
| Shared with Slipstream | Never leaves your network |
|---|---|
| The names and types of the databases the agent offers | Database server addresses and ports |
| The list of tables and columns, when someone builds a view | Database user names and passwords |
| The rows an alert or preview asks for, and nothing more | Anything else on the computer or network |
Built in, not bolted on
Who can get in
Two-step sign-in
Each person can add a code from an authenticator app to their password, and your organization can require it for its administrators or for everyone.
Sign-out after inactivity
Idle sessions end on their own: after 30 minutes for administrators and an hour for everyone else, or the times your organization chooses, with a warning first. Every session ends within 12 hours.
Roles and rights
Each person gets roles that decide what they can see and change, down to who may see the content of sent messages.
Separate organizations
Every customer is a separate organization; entities further limit each location’s data to its own people.
Audit log and sign-in history
Your administrators can see who changed what (people, roles, settings, alerts) and every sign-in and failed attempt, with the time and network address.
Your data and the system
Encrypted credentials
Database passwords and service keys you give Slipstream are stored encrypted and never shown again. Agent tokens are kept only as fingerprints.
Encrypted connections
Slipstream is served only over https, and database connections are encrypted by default.
Sensitive data guard
Check an alert before it goes live, then every run is checked too: flagged in Activity, or held for review if you prefer, with stored copies masked. A help for your own review, not a guarantee.
A record of everything sent
Activity keeps every run and message, with delivery results, so you can always answer “did it go out?”.
Told when something breaks
If a scheduled alert fails, say a database password changed, your administrators get one email naming the alert and the problem, and another when it works again.
The people you message
Opt-outs respected
Unsubscribes, STOP replies, spam reports, and bounced addresses go on a “don’t contact” list that every alert respects.
Privacy requests handled
When someone asks what you hold about them, we can find it, send them a secure copy, or delete it, while keeping them on your “don’t contact” list.
Common questions
Do we need to open a port in our firewall?
No. With the Slipstream Agent, the only traffic is outgoing https (port 443) from the agent’s computer to Slipstream. For a cloud database you can instead allow Slipstream’s fixed outgoing address and connect directly.
What does the agent need to run?
An always-on computer that can reach the database. On Windows, nothing else: the Windows download includes everything, and a guided setup asks for your database and its password, tests the login, and installs it as a service. On Linux or macOS, Node.js 22 or later, or Docker. It uses little memory. For extra resilience, install a second agent on another computer as a backup: Slipstream switches to it if the first goes offline, and your administrators are emailed when an agent is down.
What happens if our database is down?
Nothing is dropped silently. When Slipstream can’t reach the database (a network problem, a restart, an agent briefly offline), it tries the alert again after 5, 15, and 30 minutes, and the people you choose get one email if it still fails. Problems that won’t fix themselves, such as a changed password, are reported straight away instead. For planned maintenance, pause the alerts that use that database, optionally until a time when they resume by themselves; everyone sees a notice while it’s paused. When alerts resume, you choose whether each one that came due runs once now or waits for its next time. With the Slipstream Agent, a second agent on another computer can take over if the first goes offline.
Where are our database passwords kept?
With the agent, in its settings on your computer (ideally in environment variables). They are never sent to Slipstream. For direct cloud connections, they are stored encrypted in Slipstream.
Can our IT team see the queries Slipstream runs?
Yes. Slipstream writes the SQL itself from the views and criteria your team builds with clicks, and every view and criteria has a Show SQL button with the exact query: readable, or as sent to the database with values passed separately. The same query runs for previews, test runs, and alerts, whether Slipstream connects directly or through the agent. Nobody types SQL into Slipstream.
Can Slipstream write to or delete from our database?
No. It only creates queries that read data and refuses anything else, it undoes any change before a query finishes, and you give it a login that can only read.
Can Slipstream stop us sending card numbers or Social Security numbers?
It helps. Before an alert goes live, Check for sensitive data looks through what it would send for card numbers (that pass the card checksum), Social Security numbers, and attributes named like sensitive data, and shows anything it finds masked. The MASK function then sends only the last four characters. Every run is checked as it sends: it’s flagged in Activity, or, if your administrator chooses, held until someone reviews it, and the copies Slipstream keeps are stored masked. It can’t recognize everything, so it’s a safety net alongside your own review.
Do we have to give Slipstream an email password?
No. With Microsoft 365 or Google Workspace, your administrator approves Slipstream as an app that may only send email, ideally from one mailbox, and can remove that permission at any time. Mail-server (SMTP) passwords, if you use one, are stored encrypted.
Can we require two-step sign-in?
Yes. Anyone can turn it on in My account with an authenticator app such as Microsoft Authenticator or Google Authenticator, and an administrator can require it for everyone who manages users, roles, delivery, or settings, or for every person in the organization. Single sign-on with Microsoft, Google, or Okta is planned.
How is the agent updated?
On your terms. Each agent updates during a maintenance window you choose (Sundays 2–5 AM to start), as soon as a new version is out, or only when you approve it, and you can pick a time for any one update. Your administrators are emailed first with what’s new. The agent installs only versions signed by Naru Software, checking the signature itself, so not even someone who broke into Slipstream could make it run other code. Releases that need your IT never install by themselves, and one line in the agent’s settings file turns updates off entirely.
How do we cut Slipstream off?
Stop the agent, or replace or remove its token in Slipstream. Either takes effect immediately.
Have your IT team’s questions answered
We’re happy to walk your IT department through the agent and how Slipstream handles your data.