OVERVIEW
Flowframe implements enterprise-grade security measures to protect your data. This page details our comprehensive security architecture, compliance certifications, and best practices.
Key Security Highlights:
- š Dual Architecture: Secure server-based DB queries + client-side file processing
- š End-to-End Encryption: TLS 1.3 in transit, AES-256 at rest
- šŖšŗ EU Data Residency: Hosted in London, UK (GDPR-compliant)
- ā
Zero Trust Architecture: Principle of least privilege enforced
- š Compliance: GDPR-compliant
- šļø Immediate Deletion: Uploaded files deleted instantly on request (no backups)
TABLE OF CONTENTS
- Architecture Security
- Encryption
- Infrastructure and Hosting
- Access Controls
- Application Security
- Network Security
- Personnel Security
- Data Backup and Recovery
- Incident Response
- Compliance and Certifications
- Third-Party Security
- Security Monitoring
- Vulnerability Management
- Security Best Practices for Customers
- Security Roadmap
- Reporting Security Issues
1. ARCHITECTURE SECURITY
1.1 Dual Processing Architecture
Flowframe uses two distinct architectures for maximum security and flexibility:
Architecture 1: Cloud Database Connections
For cloud databases (PostgreSQL, MySQL, SQL Server, etc.):
Your Database ā Flowframe Servers ā Secure Connection ā Query Execution ā Results to Browser
ā
Encrypted transit (TLS 1.3)
ā
Results transient (not stored)
ā ļø Results pass through our servers
ā
Enhanced security & firewall compatibility
Security Measures:
- Database credentials encrypted end-to-end
- Queries executed through secure proxy servers
- Results transmitted but NOT stored (transient only)
- Connection pooling with strict isolation
- IP whitelisting support for customer firewalls
Why Through Servers:
- ā
Secure tunneling and connection management
- ā
Firewall compatibility (easier to whitelist Flowframe IPs)
- ā
Connection pooling for team collaboration
- ā
Credential rotation and management
Architecture 2: File Uploads
For uploaded files (CSV, Parquet, JSON):
Your File ā Client-Side Processing (DuckDB WASM) ā Encrypted Storage (DigitalOcean Spaces)
ā
Processed in browser first
ā
AES-256 encryption at rest
ā ļø Flowframe has access (Data Processor)
ā
Immediate deletion on request
Security Measures:
- Files processed client-side with DuckDB WASM before upload
- Encrypted with AES-256 server-side encryption in DigitalOcean Spaces
- Stored in London, UK datacenter (GDPR-compliant)
- No backups created (primary storage only)
- Immediate deletion when user requests removal
- Access strictly limited to authorized personnel
1.2 What We Store vs. What Stays Local
| Data Type | Storage Location | Encryption | Can Flowframe Access? | Duration |
|---|
| Query results (cloud DB) | Browser cache only | N/A (transient on servers) | ā ļø Transient access during transmission | Not stored |
| Query results (file uploads) | Browser cache only | N/A | ā NO - client-side only | Session only |
| Uploaded file contents | DigitalOcean Spaces | AES-256 at rest | ā ļø YES - we are Data Processor | Until deletion |
| Database credentials | Browser + Servers | AES-256 + TLS 1.3 | ā ļø YES - encrypted at rest | Until you remove connection |
| SQL query text | Our servers | AES-256 at rest | ā
YES - for version history | Until project deletion |
| Metadata descriptions | Our servers | AES-256 at rest | ā
YES - for AI and collaboration | Until project deletion |
| Account credentials | Our servers | Bcrypt hashed | ā
YES - but password hashed (we can't see plaintext) | Until account deletion |
| Collaboration state | Our servers | AES-256 at rest | ā
YES - for real-time editing | Until project deletion |
Key Security Principles:
- Query Results: Not stored on servers (cloud DB queries pass through transiently)
- File Uploads: Encrypted at rest, access controlled and logged
- Database Credentials: Encrypted end-to-end, never logged
- Zero Backups: Uploaded files have no backup copies (immediate deletion)
1.3 DuckDB WASM Security (File Processing)
DuckDB WebAssembly (WASM) executes file queries in your browser:
- ā
Sandboxed environment (browser security sandbox)
- ā
No network access for query execution
- ā
Cannot access your filesystem (except browser local storage)
- ā
Memory isolation from other browser tabs
- ā
Open-source (MIT License) - auditable by security researchers
- ā
Processing happens before files are encrypted and uploaded
1.4 Benefits of Our Architecture
Security Benefits:
- š Defense in Depth: Multiple layers of encryption and access control
- š”ļø Minimal Data Retention: Query results not stored, files deleted immediately on request
- š« Breach Mitigation: Query results transient (not affected by server breach)
- š Encryption Everywhere: TLS 1.3 in transit, AES-256 at rest
- š Access Logging: All access to uploaded files logged and monitored
- šļø Right to Erasure: Immediate deletion (no backup recovery period)
Performance Benefits:
- ā” Faster File Analysis: Client-side DuckDB WASM for file queries
- š Persistent Storage: Files available across sessions (unlike pure client-side)
- š° Cost-Effective: Secure cloud storage with no backup overhead
Compliance Benefits:
- ā
GDPR-Ready: Data Processing Agreement available, EU storage location
- ā
Right to Deletion: Immediate file deletion (no recovery period)
- ā
Data Portability: Export your files anytime
- ā
Transparency: Clear documentation of what we store and access
2. ENCRYPTION
2.1 Data in Transit
All data transmitted between your browser and Flowframe servers is encrypted:
| Protocol | Version | Cipher Suites | Certificate |
|---|
| TLS | 1.3 (latest) | AES-256-GCM, ChaCha20-Poly1305 | SHA-256 signed from trusted CA |
| HTTPS | Forced (HSTS enabled) | Perfect Forward Secrecy | 2048-bit RSA / ECC |
| WebSocket | Secure (WSS) | TLS 1.3 | Same as HTTPS |
Features:
- ā
Perfect Forward Secrecy: Past sessions protected even if keys compromised
- ā
HSTS (HTTP Strict Transport Security): Prevents downgrade attacks
- ā
Certificate Pinning: Mobile apps (planned) pin certificates for extra security
- ā
No Downgrade: TLS 1.0, 1.1, 1.2 disabled (only TLS 1.3 allowed)
Testing: Qualys SSL Labs Grade: A+ (target)
2.2 Data at Rest
All data stored on our servers is encrypted:
| Data Type | Encryption Algorithm | Key Management | Key Rotation |
|---|
| Database (PostgreSQL) | AES-256-CBC | PostgreSQL pgcrypto | Annually |
| Backups | AES-256-GCM | Separate encryption keys | Annually |
| File Storage | AES-256 | DigitalOcean Block Storage Encryption | Managed by DO |
| Logs | AES-256 | Application-level encryption | Quarterly |
Key Management:
- š Encryption keys stored separately from encrypted data
- š Key rotation: Annual rotation for database encryption keys
- š Access controls: Only authorized personnel can access keys
- š Audit logging: All key access logged
2.3 Password Security
Your password is NEVER stored in plaintext:
| Step | Method | Security Level |
|---|
| Hashing Algorithm | Bcrypt | Industry standard |
| Salt | Randomly generated per password | Unique per user |
| Rounds | 12 (cost factor) | ~300ms to hash (slows brute force) |
| Storage | Hashed password only | We cannot see your plaintext password |
What this means:
- ā
If our database is compromised, attackers get hashed passwords only
- ā
Bcrypt is designed to be slow - makes brute force attacks impractical
- ā
Salt prevents rainbow table attacks - each password hashed differently
- ā
We cannot recover your password - only reset it
Password Requirements:
- Minimum 8 characters (12+ strongly recommended)
- Complexity: Mix of uppercase, lowercase, numbers recommended
- No maximum length (up to 72 characters due to bcrypt limit)
- No password expiration (industry best practice - forced rotation leads to weaker passwords)
2.4 Database Credential Encryption
Database credentials are encrypted end-to-end:
| Storage Location | Encryption Method | Protection | Accessibility |
|---|
| In Transit (Browser to Servers) | TLS 1.3 (AES-256-GCM) | Perfect Forward Secrecy | Encrypted during transmission |
| At Rest (Flowframe Servers) | AES-256 | Encrypted in database | Flowframe (encrypted, access logged) |
Protection:
- š End-to-End Encryption - encrypted in transit via TLS 1.3
- š Encrypted at Rest - stored encrypted on Flowframe servers with AES-256
- š Access Logged - all credential access monitored and logged
- š Secure Transmission - never sent unencrypted
Recommendations:
- ā
Use read-only database credentials when possible
- ā
Enable database firewall / IP whitelisting
- ā
Rotate credentials every 90 days
- ā
Don't reuse database passwords
3. INFRASTRUCTURE AND HOSTING
3.1 Hosting Provider
Primary Host: DigitalOcean
Data Center: LON1 (London, United Kingdom)
Reason: EU data residency for GDPR compliance, low latency for European customers
| Resource | Configuration | Security Features |
|---|
| Compute | Managed Kubernetes (DOKS) | Auto-patching, isolated nodes |
| Database | Managed PostgreSQL | Automated backups, encryption, VPC isolation |
| Storage | Block Storage Volumes | AES-256 encryption, redundancy |
| Networking | Virtual Private Cloud (VPC) | Isolated network, firewall rules |
3.2 Infrastructure Security
Network Isolation:
- ā
VPC (Virtual Private Cloud): Database servers in private network with NO public internet access
- ā
Jump hosts required for admin access (not direct SSH to production)
- ā
Bastion host with MFA for privileged access
- ā
Private DNS for internal service discovery
Compute Security:
- ā
Container security: Docker images scanned for vulnerabilities
- ā
Minimal base images: Alpine Linux (smaller attack surface)
- ā
Immutable infrastructure: Containers replaced, not patched (prevents configuration drift)
- ā
Resource limits: CPU/memory limits prevent resource exhaustion attacks
Database Security:
- ā
No public IP: Database accessible only from app servers in VPC
- ā
Firewall rules: Only specific ports open
- ā
SSL/TLS required: Client connections must use TLS
3.3 Physical Security (Data Center)
DigitalOcean LON1 Data Center features:
- š¢ 24/7 Security Guards: On-site personnel
- š· Surveillance Cameras: CCTV monitoring all access points
- š Biometric Access Control: Fingerprint/retina scanners
- šŖ Mantrap Entrances: Dual-door access (one door must close before other opens)
- š„ Fire Suppression: Advanced fire detection and suppression systems
- ā” Uninterruptible Power Supply (UPS): Backup generators for power outages
- āļø Climate Control: HVAC systems maintain optimal temperature
Compliance: DigitalOcean data centers are ISO 27001, SOC 2 Type II certified.
3.4 Geographic Redundancy (Planned)
Current (Beta):
- Single region: London, UK (LON1)
Planned (Post-GA):
- Secondary region: New York (NYC3)
- Cross-region replication: Real-time database replication
- Automatic failover: If primary region fails, automatic switch to secondary
- Enterprise customers: Choose primary region (US, EU, APAC)
4. ACCESS CONTROLS
4.1 User Authentication
Login Security:
- ā
Email + Password: Industry-standard authentication
- ā
Account Lockout: After 5 failed login attempts (30-minute lockout)
- ā
Session Timeout: Automatic logout after 24 hours of inactivity
- ā
Secure Session Cookies: HTTP-only, SameSite=Lax, Secure flag
- ā
Two-Factor Authentication (2FA): TOTP-based MFA (planned Q2 2026)
Planned Enhancements (Q2 2026+):
- š§ SSO / SAML: Single Sign-On for Enterprise customers
- š§ OAuth 2.0: Social login (Google, Microsoft)
- š§ Passkeys / WebAuthn: Passwordless authentication
- š§ Hardware Security Keys: FIDO2 / YubiKey support
4.2 Role-Based Access Control (RBAC)
Workspace Roles:
| Role | Permissions | Use Case |
|---|
| Owner | Full control (manage billing, delete workspace, all admin functions) | Workspace creator, billing admin |
| Admin | Manage users, projects, settings (cannot delete workspace or change billing) | Team manager |
| Developer | Create/edit projects, run queries, collaborate | Data analysts, engineers |
| Explorer | View projects, run read-only queries, cannot edit | Stakeholders, product managers |
| Viewer | View projects only, cannot run queries or edit | Executives, external stakeholders |
Permission Matrix:
| Action | Owner | Admin | Developer | Explorer | Viewer |
|---|
| Create projects | ā
| ā
| ā
| ā | ā |
| Edit projects | ā
| ā
| ā
| ā | ā |
| Delete projects | ā
| ā
| ā
| ā | ā |
| Run queries | ā
| ā
| ā
| ā
(read-only) | ā |
| View projects | ā
| ā
| ā
| ā
| ā
|
| Add team members | ā
| ā
| ā | ā | ā |
| Manage billing | ā
| ā | ā | ā | ā |
| Delete workspace | ā
| ā | ā | ā | ā |
4.3 Administrative Access (Flowframe Personnel)
Principle of Least Privilege:
- š Only 2 senior engineers have production access
- š No customer support access to production systems
- š Temporary elevated access for debugging (time-limited, logged)
- š MFA required for all access
- š Audit logging of all admin actions
Access Tiers:
| Tier | Access Level | Who | MFA Required | Audit Logging |
|---|
| Level 1 - Support | Read-only account info (email, name, billing status) | Customer support | ā
Yes | ā
Yes |
| Level 2 - Engineering | Anonymized logs, debugging data (no PII) | Engineers | ā
Yes | ā
Yes |
| Level 3 - Admin | Full production access (emergency only) | CTO, Lead Engineer | ā
Yes | ā
Yes |
Limited Access to Your Data:
- ā Flowframe personnel cannot access your business data (query results are transient, not stored)
- ā Even with admin access, we only see SQL query text, not query results
- ā ļø Database credentials are encrypted at rest - access restricted to authorized personnel only and fully logged
4.4 API Security
API Authentication:
- š API Keys: Long, randomly generated keys (256-bit entropy)
- š Key Rotation: Customers can rotate keys anytime
- š Scoped Permissions: Keys can be scoped to specific workspaces/projects
- š Rate Limiting: 100 req/min (Free), 1,000 req/min (Pro), custom (Enterprise)
API Security Features:
- ā
HTTPS Only: All API requests require TLS
- ā
CORS Policies: Strict Cross-Origin Resource Sharing rules
- ā
Request Signing: HMAC request signatures (Enterprise API)
- ā
IP Whitelisting: Enterprise customers can whitelist IPs
- ā
OAuth 2.0: Planned for third-party integrations (Q2 2026)
5. APPLICATION SECURITY
5.1 Secure Development Lifecycle
Development Practices:
- ā
Code Reviews: All code reviewed before merging (2-person approval)
- ā
Security Training: Annual secure coding training for all engineers
- ā
Dependency Scanning: Automated scanning for vulnerable dependencies (GitHub Dependabot)
- ā
Static Analysis: Automated code security scanning (ESLint security rules, Semgrep)
- ā
Secret Scanning: Pre-commit hooks prevent accidental commit of secrets
- ā
CI/CD Security: Secure build pipelines with signed artifacts
Deployment:
- ā
Staging Environment: All changes tested in staging before production
- ā
Blue-Green Deployment: Zero-downtime deployments
- ā
Rollback Capability: Can revert to previous version within minutes
- ā
Canary Deployments: Gradual rollout to subset of users first
5.2 OWASP Top 10 Protection
Flowframe implements protections against OWASP Top 10 vulnerabilities:
| Vulnerability | Protection |
|---|
| 1. Broken Access Control | RBAC, server-side permission checks, deny-by-default |
| 2. Cryptographic Failures | TLS 1.3, AES-256, bcrypt, no sensitive data in logs |
| 3. Injection | Parameterized queries, input validation, prepared statements |
| 4. Insecure Design | Threat modeling, security architecture reviews |
| 5. Security Misconfiguration | Infrastructure-as-code, automated security configs |
| 6. Vulnerable Components | Dependency scanning, automated updates, SCA tools |
| 7. Authentication Failures | Bcrypt, MFA (planned), session management, account lockout |
| 8. Software/Data Integrity | Code signing, SRI for CDN assets, immutable infrastructure |
| 9. Logging/Monitoring Failures | Centralized logging, real-time alerts, SIEM (planned) |
| 10. SSRF | Whitelist allowed domains, network isolation, input validation |
All user input is validated:
- ā
Server-Side Validation: Never trust client-side validation alone
- ā
Whitelist Approach: Allow known-good input, reject everything else
- ā
Length Limits: All text fields have maximum lengths
- ā
Type Checking: Strict type validation (TypeScript on frontend, backend)
- ā
SQL Injection Prevention: Parameterized queries, prepared statements
- ā
XSS Prevention: Output encoding, Content Security Policy (CSP)
- ā
CSRF Protection: Anti-CSRF tokens on all state-changing requests
Content Security Policy (CSP):
Content-Security-Policy:
default-src 'self';
script-src 'self' 'wasm-unsafe-eval';
style-src 'self' 'unsafe-inline';
img-src 'self' data: https:;
connect-src 'self' https://api.flowframe.io;
frame-ancestors 'none';
base-uri 'self';
form-action 'self';
XSS Protection:
- ā
React Auto-Escaping: React escapes output by default
- ā
DOMPurify: Sanitize HTML before rendering (if rendering user HTML)
- ā
CSP: Blocks inline scripts, restricts script sources
6. NETWORK SECURITY
6.1 Firewalls
Cloud Firewall (DigitalOcean):
- ā
Inbound Rules: Only required ports are open
- ā
Outbound Rules: Allow only necessary outbound connections
- ā
Default Deny: All traffic denied by default, explicitly allow only necessary
- ā
Geo-Blocking: Block traffic from high-risk countries (optional for Enterprise)
Web Application Firewall (WAF):
- ā
OWASP ModSecurity Rules: Core rule set for common attacks
- ā
Rate Limiting: Limit requests per IP (prevents brute force, DDoS)
- ā
IP Reputation: Block known malicious IPs
- ā
Bot Detection: Identify and block malicious bots
6.2 DDoS Protection
Protection Layers:
- š”ļø Layer 3/4 DDoS: DigitalOcean Cloud Firewall (automatic mitigation)
- š”ļø Layer 7 DDoS: Application-level rate limiting, CAPTCHA challenges
- š”ļø CDN: Cloudflare (planned) for edge caching and DDoS mitigation
Mitigation Strategies:
- ā
Rate Limiting: Per-IP, per-user, per-API key
- ā
Auto-Scaling: Kubernetes auto-scaling for traffic spikes
- ā
Traffic Analysis: Real-time monitoring for anomalous traffic patterns
6.3 Network Segmentation
VPC Segmentation:
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
ā Internet ā
āāāāāāāāāāāāāāāāāā¬āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
ā
āāāāāāāāā¼āāāāāāāā
ā Load Balancer ā (Public IP)
āāāāāāāāā¬āāāāāāāā
ā
āāāāāāāāāāāāāā¼āāāāāāāāāāāāā
ā Application Tier (VPC) ā (Private IPs)
ā - Web Servers ā
ā - API Servers ā
āāāāāāāāāāāāāā¬āāāāāāāāāāāāā
ā
āāāāāāāāāāāā¼āāāāāāāāāāā
ā Database Tier (VPC) ā (Private IPs, No Internet)
ā - PostgreSQL ā
ā - Redis Cache ā
āāāāāāāāāāāāāāāāāāāāāāā
Benefits:
- š Database Isolation: No public internet access to database
- š Lateral Movement Prevention: If app server compromised, limited access to database
- š Blast Radius Reduction: Compromise contained to one tier
7. PERSONNEL SECURITY
7.1 Background Checks
For all employees with access to production systems or customer data:
- ā
Employment Verification: Confirm previous employment
- ā
Criminal Background Check: Check for criminal history
- ā
Reference Checks: Contact provided references
Timing: Before granting production access
7.2 Confidentiality Agreements
All personnel sign:
- ā
Employment Agreements: Include data protection and confidentiality clauses
- ā
Non-Disclosure Agreements (NDAs): For contractors and vendors
- ā
GDPR Confidentiality Commitments: Specific GDPR obligations
Obligations Survive: Confidentiality obligations continue after employment ends
7.3 Security Training
Mandatory Training:
- š Onboarding Security Training: All new employees (within first week)
- š Annual GDPR Training: All personnel with data access (yearly refresh)
- š Annual Security Awareness: Phishing, social engineering, password security
- š Incident Response Training: How to report and respond to security incidents
Phishing Simulations:
- š£ Quarterly Simulations: Test employee awareness
- š£ Targeted Training: Employees who click receive additional training
7.4 Access Termination
When employee leaves or role changes:
- ā±ļø Immediate Access Revocation: Within 1 hour of termination/role change
- ā±ļø Device Return: Company devices returned or remotely wiped
- ā±ļø Password Changes: Shared passwords rotated
- ā±ļø Exit Interview: Review confidentiality obligations
8. DATA BACKUP AND RECOVERY
8.1 Backup Strategy
Automated Backups:
| Type | Frequency | Retention | Encryption | Location |
|---|
| Incremental | Daily (2 AM UTC) | 30 days | AES-256 | Same region (London) |
| Full Backup | Weekly (Sunday) | 30 days | AES-256 | Same region (London) |
| Database Snapshots | Before every deployment | 7 days | AES-256 | Same region |
What's Backed Up:
- ā
Database (PostgreSQL) - Account data, metadata, collaboration state
- ā
User-uploaded configurations
- ā
Application configurations
What's NOT Backed Up:
- ā Your actual business data (stays in your database)
- ā Uploaded files (deleted immediately, no backups created)
- ā Query results (transient only, not stored)
8.2 Disaster Recovery
Recovery Objectives:
- RPO (Recovery Point Objective): Maximum 24 hours of data loss
- RTO (Recovery Time Objective): Service restoration within 4 hours
Disaster Scenarios:
| Scenario | Recovery Plan | RTO |
|---|
| Single server failure | Kubernetes auto-replaces | < 5 minutes |
| Database corruption | Restore from latest backup | < 1 hour |
| Data center failure | Manual failover to secondary region (planned) | < 4 hours |
| Complete compromise | Restore from immutable backups | < 4 hours |
Backup Testing:
- ā
Quarterly Restore Tests: Verify backups can be restored successfully
- ā
Annual DR Drills: Full disaster recovery simulation
8.3 Backup Security
Protection:
- š Immutable Backups: Write-once-read-many (WORM) storage prevents ransomware
- š Separate Credentials: Backup access uses different credentials than production
- š MFA Required: Multi-factor authentication for backup access
- š Audit Logging: All backup access logged
8.4 Customer Responsibility
You are responsible for:
- ā
Backing up your own database (we don't have access to it)
- ā
Exporting important projects regularly
- ā
Maintaining long-term archives (we only keep 30 days)
How to Backup:
- Account Settings ā Export Data (JSON format)
- Download all projects before account deletion
- Use API (Enterprise) for automated backups
9. INCIDENT RESPONSE
9.1 Incident Response Plan
Detection ā Containment ā Investigation ā Remediation ā Recovery ā Lessons Learned
9.2 Breach Notification
If a personal data breach occurs:
| Timeline | Action |
|---|
| Within 1 hour | Detect and confirm breach |
| Within 4 hours | Contain breach, prevent further unauthorized access |
| Within 24 hours | Assess scope and impact |
| Within 72 hours | Notify affected customers (GDPR requirement) |
| Within 7 days | Provide detailed incident report |
What We'll Tell You:
- ā
Nature of the breach (what data was affected)
- ā
Likely consequences
- ā
Measures taken to address the breach
- ā
Measures you can take to protect yourself
- ā
Contact point for further information
9.3 Incident Classification
| Severity | Criteria | Response Time | Escalation |
|---|
| Critical (P1) | Active breach, data exfiltration | < 15 minutes | CEO, CTO, all hands |
| High (P2) | Potential breach, security vulnerability exploited | < 1 hour | CTO, Lead Engineer |
| Medium (P3) | Security event, no data exposure | < 4 hours | Engineering Lead |
| Low (P4) | Minor security issue, no immediate risk | < 24 hours | Assigned engineer |
9.4 Communication
During incident:
- ā
Status Page: Real-time updates at status.flowframe.io (planned)
- ā
Email Notifications: Affected customers notified
- ā
Transparency: Honest communication about what happened
Post-Incident:
- ā
Incident Report: Detailed post-mortem within 7 days
- ā
Lessons Learned: What we learned and how we'll prevent recurrence
- ā
Remediation Plan: Steps taken to fix root cause
10. COMPLIANCE AND CERTIFICATIONS
10.1 Current Compliance Status
| Framework | Status | Details |
|---|
| GDPR | ā
Compliant | Data Processing Agreement available, EU hosting, SCCs in place |
| CCPA | ā
Compliant | California Consumer Privacy Act obligations met |
| UK GDPR | ā
Compliant | UK data protection law compliance |
| ePrivacy Directive | ā
Compliant | Cookie consent, email marketing compliance |
| ISO 27001 | š§ Under evaluation | Information Security Management System |
| PCI DSS | ā
Via Stripe | We don't handle card data directly (Stripe is PCI DSS Level 1) |
10.2 Data Protection Impact Assessment (DPIA)
Customers may require DPIA if:
- Processing likely to result in high risk to data subjects
- Large-scale processing of special categories of data
- Systematic monitoring of publicly accessible areas
Flowframe's Support:
- ā
Provide technical documentation for customer DPIA
- ā
Describe processing operations and purposes
- ā
Assess necessity and proportionality
- ā
Describe security measures
Request DPIA Support: support@flowframe.io
10.3 Sub-Processor Due Diligence
All sub-processors assessed for:
- ā
Security certifications (SOC 2, ISO 27001, etc.)
- ā
Data protection compliance (GDPR, CCPA)
- ā
Incident response capabilities
- ā
Encryption and access controls
- ā
Financial stability and reputation
Sub-Processors:
- DigitalOcean: SOC 2 Type II, ISO 27001
- Stripe: PCI DSS Level 1, SOC 2, ISO 27001
- Google (Gemini, planned): SOC 2, ISO 27001, Zero Data Retention
11. THIRD-PARTY SECURITY
11.1 Sub-Processor Security Standards
All sub-processors must:
- ā
Sign Data Processing Agreements (DPAs)
- ā
Implement appropriate technical and organizational measures
- ā
Provide data breach notification procedures
- ā
Maintain security certifications (SOC 2, ISO 27001)
- ā
Comply with GDPR and applicable data protection laws
- ā
Submit to regular security audits
11.2 Vendor Risk Management
Vendor Assessment Process:
- Initial Assessment: Security questionnaire (SIG, CAIQ)
- Due Diligence: Review certifications, policies, incident history
- Contract Negotiation: Data Processing Agreement, SLAs, security requirements
- Ongoing Monitoring: Quarterly review, annual re-assessment
- Incident Tracking: Monitor for security incidents at vendor
Vendor Offboarding:
- ā
Data deletion confirmation
- ā
Access revocation
- ā
Certificate of destruction (if applicable)
11.3 Integration Security
For third-party integrations (databases, APIs):
- ā
Customer Controls: You manage your own database credentials
- ā
Read-Only Recommended: Use read-only credentials when possible
- ā
Encrypted in Transit: Credentials encrypted via TLS 1.3 during transmission
- ā
Encrypted at Rest: Credentials stored encrypted on Flowframe servers (AES-256)
12. SECURITY MONITORING
12.1 24/7 Monitoring
What We Monitor:
- š Failed Login Attempts: Alert on brute force patterns
- š Anomalous Traffic: Unusual traffic patterns, DDoS attempts
- š Error Rates: Spikes in application errors (potential attack)
- š System Performance: CPU, memory, disk usage anomalies
- š Network Traffic: Unauthorized network connections
- š Administrative Actions: All admin access logged and monitored
Alerting:
- ā ļø Real-Time Alerts: Critical events trigger immediate alerts
- ā ļø Escalation: On-call engineer notified within minutes
- ā ļø Runbooks: Documented response procedures for common incidents
12.2 Logging and Audit Trails
Comprehensive Logging:
| Log Type | Retention | Content | Encryption |
|---|
| Application Logs | 90 days | User actions, API calls | AES-256 |
| Access Logs | 90 days | Authentication, authorization | AES-256 |
| Admin Logs | 1 year | All admin actions | AES-256 |
| Security Logs | 1 year | Security events, failed logins | AES-256 |
| Audit Logs | 3 years | GDPR-related activities (DSAR, data deletion) | AES-256 |
Log Analysis:
- š Automated Analysis: Machine learning for anomaly detection (planned)
- š SIEM (Security Information and Event Management): Planned Q3 2026
- š Correlation: Correlate events across systems to detect complex attacks
12.3 Security Metrics
KPIs Tracked:
- ā±ļø Mean Time to Detect (MTTD): How quickly we detect incidents
- ā±ļø Mean Time to Respond (MTTR): How quickly we respond to incidents
- š Failed Login Rate: Indicator of brute force attempts
- š Vulnerability Remediation Time: Time to patch vulnerabilities
- šÆ Phishing Simulation Success Rate: Employee security awareness
13. VULNERABILITY MANAGEMENT
13.1 Security Updates
Patching Schedule:
- šØ Critical Vulnerabilities: Within 24 hours
- ā ļø High Severity: Within 7 days
- š Medium Severity: Within 30 days
- š Low Severity: Next quarterly maintenance window
Dependency Management:
- ā
Automated Scanning: GitHub Dependabot scans dependencies daily
- ā
Automated PRs: Dependabot creates PRs for security updates
- ā
Manual Review: All dependency updates reviewed before merging
13.2 Penetration Testing
External Penetration Testing:
- šÆ Frequency: Under evaluation for future implementation
- šÆ Scope: Full application and infrastructure penetration test
- šÆ Provider: Independent third-party security firm
- šÆ Report: Detailed findings and remediation recommendations
13.3 Vulnerability Disclosure
If you find a security vulnerability:
- Email: support@flowframe.io with subject "SECURITY VULNERABILITY REPORT"
- Provide: Description, steps to reproduce, potential impact
- Responsible Disclosure: Give us time to fix before public disclosure (90 days)
- Our Response:
- Acknowledge receipt within 24 hours
- Investigate and respond within 5 business days
- Keep you updated on remediation progress
14. SECURITY BEST PRACTICES FOR CUSTOMERS
14.1 Account Security
Recommendations:
- ā
Use a strong, unique password (12+ characters)
- ā
Enable two-factor authentication (when available Q2 2026)
- ā
Don't share credentials with anyone
- ā
Use a password manager (1Password, Bitwarden, LastPass)
- ā
Log out when using shared computers
14.2 Database Security
Recommendations:
- ā
Use read-only database credentials when possible
- ā
Create a dedicated Flowframe user with minimal permissions
- ā
Enable IP whitelisting if your database supports it
- ā
Rotate credentials every 90 days
- ā
Don't use root/admin credentials
- ā
Monitor database access logs
Database User Permissions (Example for PostgreSQL):
-- Create read-only user for Flowframe
CREATE USER flowframe_readonly WITH PASSWORD 'strong-password-here';
GRANT CONNECT ON DATABASE your_database TO flowframe_readonly;
GRANT USAGE ON SCHEMA public TO flowframe_readonly;
GRANT SELECT ON ALL TABLES IN SCHEMA public TO flowframe_readonly;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT SELECT ON TABLES TO flowframe_readonly;
14.3 Data Handling
Recommendations:
- ā
Only upload data you're authorized to analyze
- ā
Anonymize or redact sensitive data before public sharing
- ā
Export and backup important analyses regularly
- ā
Delete old projects you no longer need
- ā
Use RBAC to limit team member access appropriately
14.4 Team Management
Recommendations:
- ā
Remove team members who leave your organization immediately
- ā
Use role-based access (don't give everyone Owner access)
- ā
Review team member list quarterly
- ā
Audit project sharing settings periodically
- ā
Educate team on security best practices
15. SECURITY ROADMAP
15.1 Current (Q1 2026) - Beta Phase
Implemented:
- ā
TLS 1.3 encryption
- ā
AES-256 encryption at rest
- ā
Client-side data processing (DuckDB WASM)
- ā
Role-based access control
- ā
EU data residency (London, UK)
- ā
GDPR compliance
- ā
Bcrypt password hashing
- ā
VPC network isolation
- ā
Automated backups
- ā
24/7 security monitoring
15.2 Coming Q2 2026 - General Availability
Planned:
- š§ Two-Factor Authentication (2FA): TOTP-based MFA
- š§ SSO / SAML Integration (Enterprise): Single Sign-On
- š§ Bug Bounty Program: Public vulnerability disclosure
- š§ Security Vulnerability Disclosure Policy
- š§ Enhanced Audit Logs: More granular activity tracking
15.3 Future (H2 2026 & Beyond)
Roadmap:
- š® IP Whitelisting: For API access
- š® Advanced Audit Logs: SIEM integration
- š® Data Loss Prevention (DLP): Automated sensitive data detection
- š® Customer-Managed Encryption Keys (CMEK): Bring your own encryption keys
- š® Hardware Security Keys: FIDO2/YubiKey support
- š® Passkeys / WebAuthn: Passwordless authentication
16. REPORTING SECURITY ISSUES
16.1 Vulnerability Reporting
Found a security vulnerability?
Contact: support@flowframe.io
Subject: "SECURITY VULNERABILITY REPORT"
Please Include:
- ā
Description of the vulnerability
- ā
Steps to reproduce
- ā
Potential impact
- ā
Your contact information
- ā
Any proof-of-concept code (if applicable)
Our Response:
- ā
Acknowledge receipt within 24 hours
- ā
Investigate and respond within 5 business days
- ā
Keep you updated on remediation progress
- ā
Credit you in security advisory (if you wish)
- ā
Potential bounty reward (once program launches Q2 2026)
16.2 Responsible Disclosure
We ask that you:
- ā
Give us 90 days to fix the vulnerability before public disclosure
- ā
Don't exploit the vulnerability beyond proof-of-concept
- ā
Don't access other users' data
- ā
Don't perform DoS attacks or service disruptions
- ā
Don't publicly disclose until we've fixed the issue
In return, we promise:
- ā
No legal action against security researchers acting in good faith
- ā
Transparent communication and updates
- ā
Public recognition (if you wish)
- ā
Fair bounty rewards (once program launches)
SUMMARY
Flowframe implements enterprise-grade security:
- š Dual Architecture: Secure cloud DB queries + encrypted file storage (AES-256)
- š End-to-End Encryption: TLS 1.3 in transit, AES-256 at rest
- šŖšŗ EU Data Residency: GDPR-compliant hosting in London, UK
- ā
Compliance: GDPR, CCPA
- š”ļø Zero Trust Architecture: Least privilege, MFA, network isolation
- š Audit Ready: Comprehensive logging, incident response, breach notification
- šļø Immediate Deletion: Files deleted instantly on request (no backups)
Questions? Contact us at support@flowframe.io
Last Updated: January 10, 2026
Version: 1.0
Next Review: April 2026
For more security information:
We take security seriously. If you have suggestions for improving our security posture, we'd love to hear from you at support@flowframe.io.
END OF SECURITY PAGE