01Enterprise Trust & Compliance Posture
SadaSend is engineered from the ground up to satisfy modern enterprise security requirements, data protection standards, and zero-trust deliverability practices.
Our infrastructure is hosted in ISO 27001-certified datacenter facilities in Germany (EU), run on renewable energy.
- Certifications & Assurance Roadmap: the underlying datacenter facilities in Germany (EU) are ISO 27001-certified. SadaSend itself does not currently hold SOC 2 or HIPAA certification and does not sign HIPAA business associate agreements; do not send protected health information. We are preparing for an external third-party penetration test and SOC 2 Type I readiness assessment on our enterprise compliance roadmap.
- PostgreSQL Row-Level Security (RLS): Every database query runs within a strict tenant-isolated context, preventing cross-tenant data leakage.
- Tamper-Proof Audit Logging: Immutable system audit logs strictly forbid UPDATE and DELETE operations at the database engine level, guaranteeing an unalterable record of every staff action, including every read of a message body.
- Container Sandbox Isolation: Production containers run under dedicated unprivileged system users with capability dropping, read-only root filesystems, PID bounds, and explicit memory limits.
- Data Isolation: tenant-isolated queries, API keys stored only as hashes, TLS 1.2 or higher in transit, and provider-managed encryption at rest.
- Enterprise Contracting: Bespoke Master Service Agreements (MSAs), custom data processing agreements (DPAs), and direct procurement support for scale customers.
02Encryption
- All API and dashboard traffic is served over TLS 1.2 or higher. HTTP requests are redirected, never served.
- SMTP submission on port 587 requires STARTTLS. We do not accept unencrypted authentication.
- Outbound mail uses opportunistic TLS to receiving servers, which is the strongest guarantee the email protocol allows.
- Data at rest — database and object storage — is encrypted using our infrastructure providers’ managed encryption.
03Credentials
- API keys are stored as hashes. We cannot show you a key again after you create it, because we do not have it.
- Every key carries an explicit scope set, its own rate limit, an optional recipient allowlist, and an optional approval mode. A leaked key does strictly less damage than a leaked key elsewhere.
- API keys support IP allowlists (CIDR restrictions) and customizable expiration dates to limit attack surfaces.
- Out-of-Band OTP Challenge: Generating new live API keys requires a single-use 6-digit cryptographic OTP verification challenge dispatched to the verified account owner, neutralizing session hijacking threats.
- Hardware Passkeys & WebAuthn for Staff: Administrative operations enforce FIDO2 WebAuthn hardware security keys for sensitive write operations, message inspection, and customer security hold releases.
- DKIM private keys are generated per domain, encrypted at rest, and never exposed through the API or dashboard.
- There are no passwords to steal: you sign in with a one-time code or link sent to your email, or with GitHub or Google. Authenticator-app two-factor authentication is available on all plans, including Free.
04Access control
Access to production systems is limited to engineers who need it, requires two-factor authentication, and is logged. We do not read customer message bodies as a matter of routine — only when you ask us to investigate a specific message, or where we are legally compelled.
05Abuse prevention
A shared sending platform is only as trustworthy as its worst customer, so abuse controls are a security feature rather than an operational afterthought.
- New accounts start with a low daily limit that rises automatically over the first weeks while bounce and complaint rates stay healthy, and stops rising when they do not.
- Bounce and complaint rates are evaluated per account continuously, with automatic throttling and suspension.
- Sending from an unverified domain is refused outright, which prevents most impersonation attempts.
- Suppression is enforced when a message is accepted rather than when it is sent, so a suppressed address fails fast and visibly.
06Autonomous Agent Guardrails & Outbound DLP
Autonomous AI agents and automated workflows introduce new failure modes: runaway loops that spam users and accidentally leaked API keys in email bodies. SadaSend inspects dispatches before transmission to protect your recipients and credentials.
- Outbound DLP Secret Scanner: inspects every message for leaked private keys and API keys (AWS, OpenAI, Anthropic, Stripe, GitHub, Slack, Google, SendGrid, Resend), card numbers and US Social Security numbers, and refuses it with 403 dlp_violation.
- Runaway Loop Circuit Breaker: Automatically detects and pauses duplicate messages sent within a rolling 5-minute window, preventing automated loops from spamming inboxes or draining quota.
- Granular Key Scopes & Allowlists: Enforce hourly send limits and restrict API keys to pre-approved recipient domains (e.g. @yourcompany.com).
- Approval Mode & Slack Reviews: Hold sensitive or agent-generated emails for 1-click human verification in your dashboard or directly inside Slack before transmission.
07Reporting a vulnerability
Write to security@sadasend.com. We commit to acknowledging vulnerability reports within 24 hours, completing triage within 48 hours, and deploying patches within 24 to 72 hours for critical findings.
Our canonical RFC 9116 security contact information is published at https://sadasend.com/.well-known/security.txt.
We will not pursue legal action against researchers who act in good faith, stay within their own account’s data, avoid degrading the service for others, and give us reasonable time to fix an issue before publishing. We do not currently run a paid bounty programme, and we will say so plainly rather than imply one exists.
08Reporting abuse
If you have received unwanted mail sent through SadaSend, report it to abuse@sadasend.com with full message headers. Reports are investigated by a person, and we act on the sending account rather than asking you to unsubscribe.