Security
Security is a feature, not a page.
The same rules that keep your team safe also bind the AI assistant. Here's what that means in practice.
Tenant isolation
Every query is scoped to your workspace. Cross-workspace references are rejected on write.
AI with guardrails
The assistant can read; it can't write. Proposals are permission-checked and require human approval.
Roles & permissions
Owner, Admin, Manager, Member and Viewer roles with 28 fine-grained permissions.
Session security
Short-lived access tokens held in memory, rotating httpOnly refresh tokens and CSRF protection.
Compliance in the write path
Do-not-contact checks run before every send, on every channel, not in a nightly batch.
Auditable spend
An append-only ledger records every credit movement with who, what and balance after.
Practices
How it works under the hood.
- Data in transit
- TLS for all traffic between your browser, our API and our providers.
- Sessions
- Access tokens expire after 15 minutes and are never written to browser storage. Refresh tokens rotate on every use; reusing an old one fails.
- Passwords
- Hashed with bcrypt. Sign-in responses take the same time whether or not an email exists.
- Workspace isolation
- Every read and write is scoped to the caller's workspace. Writes that reference another workspace's records are rejected.
- AI providers
- Workspace data is sent to the model provider only to answer a request you made. It isn't used to train models.
- Outbound compliance
- Do-not-contact checks, inbox caps and credit holds run in a single send path, so no feature can skip them.
- Responsible disclosure
- Found something? Email security@autleads.example and we'll respond within two business days.
100 free credits, no card
Put your next hundred prospects on the map.
Start free with the full workflow: map search, website emails, sequences and the CRM.