Our values

Security by architecture, not by policy.

Guardrails that are built into the system, not as preferences to follow.

Vaulted credentials

Keys stay in a backend vault, never seen by the AI.

Human approval

Sending or deleting always waits for your yes.

Never trained on your data

Your messages and files train no model.

Isolated workspaces

Sandboxed per workspace, never shared.

What to expect

Your biggest concerns, addressed head-on.

The AI leaking your passwords or keys
It never holds them. Credentials are added only at request time: the AI can use your connected tools without ever seeing them.
One company's data reaching another
Every customer is fully isolated. There's no path between workspaces.
One teammate seeing another's private chats or accounts
Access is scoped to each person: everyone sees only their own.
The AI deleting, sending, or spending on its own
Anything risky pauses for approval, and only the person who triggered it can approve.
Getting mass-emailed on your behalf
Bulk email is switched off permanently.

The architecture

How our security works.

Every guarantee above traces back to a specific control: each one enforced by the system, never left to the AI.

Data scoped per user

Connected accounts, files, and messages stay tied to the person who created them, checked on our servers before the AI sees anything.

Hardened against prompt injection

AI can be tricked by hidden instructions in the content it reads. We clean everything coming in and flag anything trying to sneak data out.

Secrets can't leak outbound

Everything Superpal sends is scanned for passwords and keys. If something sensitive shows up, it's blocked before it leaves.

Costs can't run away

Built-in spending limits stop runaway usage, and an emergency switch can cut off paid activity in seconds.

Locked-down access

Customers sign in through Slack. Internal admin access needs two-factor, is limited to a few people, and every action is logged.

Full audit trail

Every action is recorded: who did what, which tool, and when. Admins get a full activity view across the workspace.

Where we stand

More secure every day.

The honest state of our compliance, kept current.

  • SOC 2 Type IIPlanned 2026
  • ISO 27001Planned 2026
  • Data Processing AgreementOn request

Responsible disclosure

Found something? Tell us.

If you spot a vulnerability, report it directly and we'll act fast. We credit researchers who help keep Superpal safe.

support@superpal.ai

FAQ

Ask the uncomfortable ones.

The questions you'd actually want answered before letting an AI into your workspace. Here they are, answered straight.

You
Can you read our DMs?
Superpalapp

Get started

Your next hire is one message away.

Add Superpal to Slack, delegate your first task, and get finished work back today.

$100 in free credits · no credit card