Encryption & Keys
TLS, AES-256-GCM, organization-managed and end-to-end encrypted rooms.
Security
Authentication, encryption, access control, and monitoring.
Detailed controls are organized by topic. Start with the overview below or jump to a focused page.
TLS, AES-256-GCM, organization-managed and end-to-end encrypted rooms.
Passkeys, SSO, RBAC, sessions, and SCIM provisioning.
Environment separation, ingress hardening, and operator access.
CSRF, CSP, dependency scanning, and secure SDLC practices.
Backups, monitoring, and incident response readiness.
Data ownership, retention, export, and subprocessors.
Access is scoped to roles, rooms, and administrative functions required for each user.
Layered controls across identity, network, application, and data layers.
Sensible defaults for authentication, session handling, and workspace permissions.
Organizations configure providers, encryption modes, retention, and access policies.
Algorithms coThink uses and where they apply.
| Control | Algorithm / Standard | Application |
|---|---|---|
| Transport encryption | TLS 1.2+ (TLS 1.3 preferred) | All browser, API, and webhook traffic between clients and coThink services |
| Symmetric encryption | AES-256-GCM | Message content, provider API keys, organization chat keys, and other designated secrets at rest |
| Password hashing | bcrypt (cost factor 10) | Local account passwords; plaintext passwords are never stored |
| Key wrapping (E2EE rooms) | RSA-OAEP with SHA-256 (4096-bit keys) | Per-message AES content keys wrapped for each recipient public key |
| Authentication (passkeys) | WebAuthn / FIDO2 | Passwordless sign-in and multi-factor verification |
| Integrity | GCM authentication tags | Authenticated encryption for stored secrets and encrypted message envelopes |
Permissions are enforced at the organization, workspace, and room level. Administrative capabilities are separated from standard collaboration access.
Room membership, invitation flows, and content visibility respect configured workspace and room policies.
Authorized administrators can configure room encryption policies, including locking rooms to organization-managed or end-to-end encrypted modes.
AI provider API keys are stored encrypted and are only decrypted server-side when required to fulfill authorized inference requests.
All connections between browsers, mobile clients, and coThink APIs use TLS. Production deployments enforce HTTPS and HTTP Strict Transport Security (HSTS) on the application tier.
Outbound connections to customer-configured AI providers, calendar systems, and other integrations are made over TLS. Customers should verify provider endpoint security as part of their vendor review.
Designated sensitive data is encrypted with AES-256-GCM before persistence. Each encryption operation uses a unique 96-bit initialization vector (IV) and an authentication tag to detect tampering.
Database and persistent storage volumes rely on provider-managed encryption at the infrastructure layer, in addition to application-layer encryption for designated secrets and encrypted content.
Platform encryption keys are held in secure environment configuration and are not stored alongside encrypted data.
coThink supports two optional room encryption modes. Both use AES-256-GCM for message content; they differ in who holds the decryption keys.
ct-org-v1)Suitable for organizations that want encrypted storage with server-side processing (AI inference, search, exports) on authorized content.
ct-e2ee-v1)For teams that need client-side keys. If private keys are lost, coThink cannot recover E2EE content.
Customer-supplied AI provider credentials are encrypted with AES-256-GCM using a platform-held encryption key before database storage. API responses never return plaintext keys.
Organization chat key material is fingerprinted (SHA-256) for identification and stored encrypted. Only authorized administrative flows can generate, import, or rotate keys.
Production secrets (encryption keys, database credentials, integration tokens) are supplied via secure environment configuration managed through infrastructure automation—not committed to source control.
Structured logging pipelines redact patterns matching API keys, tokens, passwords, encryption keys, and other sensitive fields before emission.
Security-relevant actions—including authentication events, administrative changes, encryption policy updates, and key management operations—are recorded for review.
Operational telemetry supports detection of anomalous access patterns, failed authentication attempts, and service health degradation.
Background workers, API services, and integration health are monitored to support availability and timely incident response.
Security issues can be reported through the Trust Center contact page. coThink investigates reported vulnerabilities and coordinates remediation according to severity.
Procurement FAQ · Compliance roadmap · Report a security issue →