Security & Compliance

Pingd takes security seriously. This page details our security architecture, data handling practices, and compliance posture.

Data Architecture

Encryption

• In transit: All data is encrypted using TLS 1.3
• At rest: AES-256 encryption for all stored data
• Database: Hosted on Supabase with row-level security (RLS) policies
• Connector credentials: All external connection configurations (database passwords, API keys, OAuth tokens) are encrypted before storage — credentials are never stored in plaintext

Data Isolation

Each organization's data is logically isolated:

• Row-level security ensures no cross-org data access
• API calls are scoped to the authenticated user's organization
• Agent queries only access data within the user's policy scope
• External connections are scoped per-organization — one org cannot access another's connectors

Data Residency

• Primary data storage: US (AWS us-east-1)
• Contact us for EU data residency options (Enterprise plan)

Row-Level Security (RLS)

Pingd enforces RLS at the database level on all sensitive tables. This means even if application logic has a bug, the database itself prevents unauthorized access.

Tables protected by RLS:

• copilot_api_keys — API keys scoped to the owning organization
• external_connections — Data source credentials and configurations
• external_tables — Discovered table schemas and mappings
• slack_workspace_connections — Slack workspace-to-org mappings
• agent_data_policies — Policy rules scoped to the owning organization
• external_query_logs — Query audit trail scoped per-org
• All existing tables (organizations, profiles, insights, etc.)

Every query to these tables is automatically filtered by organization_id, enforced at the Postgres level.

Access Controls

Authentication

• Email/password with bcrypt hashing
• OAuth via Google and Microsoft
• Two-factor authentication (TOTP) available
• SSO via SAML 2.0 (Enterprise plan)

API Key Authentication

For programmatic access (e.g., the Copilot API), Pingd uses API keys:

• Keys are SHA-256 hashed before storage — the raw key is shown once at creation and never stored
• Each key has a key_prefix (first 8 characters) for identification without exposing the full key
• Keys are scoped to a specific organization via org_id
• Keys support is_active toggle and optional expires_at for time-limited access
• last_used_at is tracked for auditing
• Authentication uses Bearer token in the Authorization header

API keys are managed at Admin → API Keys in the dashboard.

Authorization

Multiple layers of access control:

1. Role-based — Owner, Admin, Member roles control feature access
2. Policy-based — Data access policies control data visibility (see Data Access Policies)
3. Row-level — Database RLS enforces access at the Postgres layer
4. Org-scoped — All resources (connections, policies, keys) are isolated per organization

Session Management

• JWT-based session tokens via Supabase Auth
• Configurable session timeout
• Automatic token refresh
• Session revocation on password change

External Data Source Security

Credential Storage

When you connect an external data source (Snowflake, S3, Salesforce, Redshift, PostgreSQL), your credentials are:

• Encrypted before being stored in the connection_config_encrypted field
• Never logged or included in error messages
• Only decrypted at query time in server-side memory
• Scoped to your organization — no other org can access your connection

Connection Types & Auth Methods

Source	Authentication
Snowflake	Username/password + account identifier
Amazon S3	AWS access key + secret key
Salesforce	OAuth 2.0 with automatic token refresh
Redshift	Username/password via PostgreSQL protocol
PostgreSQL	Username/password + SSL

Query Security

• All queries pass through the policy engine before reaching your data source
• Row-level filters are injected into every query automatically
• Query logs are maintained for audit purposes in external_query_logs
• The agent never has direct database access — all queries go through the Pingd API

AI & LLM Security

Data Usage

• Your data is never used to train AI models
• Queries are processed in real-time and not retained by AI providers
• We use enterprise AI APIs with data processing agreements

Prompt Security

• Input sanitization prevents prompt injection
• System prompts are server-side only (not visible to users)
• Guardrails prevent the agent from revealing sensitive configuration

Response Filtering

• PII detection in responses
• Configurable blocked topics
• Rate limiting per user

Audit Trail

All significant actions are logged:

• User logins and logouts
• Data access queries (including which connector was used)
• Configuration changes
• Policy modifications
• API key creation and revocation
• External connection changes
• User management actions (invites, role changes, removals)

Access the audit log at Settings → Security → Audit Log.

Audit logs are:

• Immutable (cannot be modified or deleted)
• Retained for 2 years
• Exportable as CSV for compliance reviews

Compliance

SOC 2 Type II

Our architecture is designed to meet SOC 2 Type II requirements:

• Access controls and authentication
• Data encryption at rest and in transit
• Audit logging and monitoring
• Incident response procedures
• Change management processes

SOC 2 certification in progress — contact us for our current security questionnaire.

GDPR

For organizations with EU users or data:

• Data processing agreement (DPA) available
• Right to deletion supported
• Data export available
• EU data residency option (Enterprise plan)

CCPA

California Consumer Privacy Act compliance:

• Data inventory and mapping
• Deletion requests honored within 30 days
• No sale of personal information

Incident Response

In the event of a security incident:

1. Detection — Automated monitoring and alerting
2. Containment — Immediate isolation of affected systems
3. Notification — Affected organizations notified within 72 hours
4. Remediation — Root cause analysis and fix deployment
5. Review — Post-incident review and process improvement

Security Questionnaire

Need to complete a vendor security assessment? Contact security@pingd.io — we have pre-completed questionnaires for common frameworks (SIG, CAIQ, VSAQ).

Pro Tip: For enterprise security requirements (SSO, IP allowlist, custom data retention, dedicated infrastructure), contact sales@pingd.io for our Enterprise plan.

See also: Data Access Policies · Admin Guide · Organization Settings