Security for
agentic systems.

Your data stays in your region, and every agent stays on the record.

Data residency

Your data stays in your region. The EU or the US, picked at setup.

  • EU (Frankfurt) or US

Region

Where this workspace stores its data.

Data
Agent runs
Sandboxes
Agent databases

Encrypted credentials

Agents never see your keys. We add them on our servers, outside the model.

  • AES-256-GCM encryption
  • Calls can't be redirected

HubSpot · API key

••••••••••••••••

Encrypted · never shown to the model

Isolated sandboxes

Every agent works alone. Its own sandbox, files and code.

  • Storage in your region

Deal Desk

Invoice Match

Ticket Triage

Renewal Watch

Lead Research

Payroll Check

Roles

A role for everyone. Five built in, or make your own.

Role

Can manage

Billing
Budgets
Integrations
API keys
App activity
Settings

Tool access

You decide what agents can reach. Allow or block every app and action.

Ashby

Get Candidate
List Applications
Move Candidate Stage
Add Note
Archive Candidate

Single sign-on

Sign in with your identity provider. On the Enterprise plan.

  • Okta or any SAML provider
  • Required for your domain
Continue with OktaContinue with SAMLEmail and password

SSO required for acme.com

Activity records

Every tool call is recorded. Who, which agent, and the result.

14:0048 calls · 4 blocked

Move candidate14:42
Send email14:31
Create opportunity14:08

Confidential agents

Private work stays private. Every action recorded, never the content.

Create Opportunity · Payroll Check · 15:49Confidential

Input

Hidden

Output

Hidden

AI models

Only the models you approve. On or off, region by region.

  • Anthropic, OpenAI, Google, xAI, open models
  • Never used for training
Claude Opus 5.5
Claude Sonnet 5.5
GPT-5.5
Gemini 3 Pro
DeepSeek V4
GDPR

GDPR

Built to GDPR standards, so your data stays private and protected.

Ready for your security review. Our standards and documents, in one place.

GDPR

Compliant, with a DPA

SOC 2 Type II

In progress

ISO 27001

In progress

What security teams ask us.

Where is our data hosted?

In the EU (Frankfurt) or the US, chosen when you create your workspace. Your data stays there.

Can the AI model see our API keys or passwords?

No. Credentials are encrypted and added on our servers, so the model never sees them.

Can we see everything our agents did?

Yes. Every tool call is recorded, blocked ones too. Export it to CSV or your SIEM.

Can we stop agents from using certain apps?

Yes. Allow or block any app or action, for the company, a team or one person.

Is our data used to train AI models?

No. Your data is never used to train AI models.

Do you support single sign-on?

Yes, on the Enterprise plan.

What happens to our data if we leave?

A deleted workspace is permanently purged after 30 days.

Bring your security team.We'll bring the answers.