Privacy & Security
Reliant is built with privacy and security as core principles. Your code and conversations belong to you, and we’re committed to protecting them.
Data Ownership
Your data is yours. Period.
- We will never sell your data
- We will never use your data to train models across accounts
- We will never share your conversations or code with third parties
- Your data is stored locally and optionally synchronized to your private cloud storage
What Data is Stored
Locally Stored Data
Reliant stores the following data on your local machine:
Conversation History
- Messages between you and AI agents
- Tool calls and their results
- Agent transitions and state changes
- Session metadata (timestamps, titles, agents used)
Location: Application data directory (e.g., ~/Library/Application Support/Reliant/ on macOS)
Configuration
- API keys (encrypted)
- Model preferences
- MCP server configurations
Location: .reliant.json in application data directory
Project Data
- Custom agent definitions (
.reliant/agents/) - Custom taskforce configurations (
.reliant/taskforces/) - Project-specific settings
Location: Within your project’s .reliant/ directory
Optionally Synchronized Data
If you enable remote sessions (for mobile access or multi-device sync):
Session Synchronization
- Conversation history (to resume on other devices)
- Tool call records (for context continuity)
- Agent configurations (for consistent experience)
Not Synchronized:
- API keys (never leave your device)
- Local file contents (only file paths are stored)
- Git repository data
- Build artifacts or generated files
Security Modes
Standard Mode (Default)
In standard mode:
- Conversations are stored locally in SQLite database
- API keys are stored encrypted in configuration files
- Tool calls have access to your file system within project scope
- Network requests go directly to your configured AI providers
Security Mode
Enable security mode for maximum privacy:
{
"securityMode": true
}
When security mode is enabled:
- All file paths are scrubbed before sending to AI models
- Sensitive patterns (API keys, tokens, passwords) are detected and redacted
- Network requests are logged for audit
- Tool call permissions require explicit confirmation
- Model responses are scanned for potential data leakage
Note: Security mode may impact agent effectiveness as context is more limited.
Offline Mode
Work completely offline:
{
"offlineMode": true,
"providers": {
"local": {
"endpoint": "http://localhost:1234/v1"
}
}
}
In offline mode:
- Only local models are available
- No data leaves your machine
- No internet connection required
- All processing happens locally
Telemetry and Analytics
What We Collect
Reliant collects minimal telemetry to improve the product:
Usage Statistics (Anonymous):
- Feature usage counts (which tools, agents used)
- Performance metrics (response times, error rates)
- Configuration patterns (which models are popular)
Crash Reports (Optional):
- Stack traces for unhandled errors
- System information (OS version, Reliant version)
- Error context (what operation failed)
What We DON’T Collect:
- Your code or file contents
- Conversation messages
- API keys or credentials
- Personal information
- File paths or project names
Disabling Telemetry
You can completely disable telemetry:
Via Settings UI:
- Open Settings
- Navigate to Privacy section
- Toggle "Enable Telemetry" off
Via Configuration File:
{
\"telemetry\": {
\"enabled\": false,
\"crashReports\": false
}
}
Via Environment Variable:
export RELIANT_TELEMETRY_DISABLED=1
Disabling telemetry does not affect functionality - Reliant works exactly the same.
Data Deletion
Deleting Local Data
Delete All Application Data (macOS):
rm -rf ~/Library/Application\ Support/Reliant/
For other platforms, delete the application data directory in the equivalent location.
Delete Specific Sessions:
- In Reliant UI: Right-click session → Delete
- Or delete the SQLite database in the application data directory
Reliant will recreate necessary directories and databases on next launch.
Delete Project Data:
# In your project directory
rm -rf .reliant/
This removes custom agents, taskforces, and project-specific configuration.
Deleting Remote Data
If you’ve enabled remote sync:
- Go to Settings → Privacy
- Click "Delete All Remote Data"
- Confirm deletion
- All synced conversations are permanently deleted from cloud storage
Important: Deletion is permanent and cannot be undone.
API Keys and Credentials
How API Keys Are Stored
Encrypted Storage:
- API keys are encrypted using AES-256
- Encryption key is derived from system-specific data
- Keys are never stored in plain text
Location:
- Configuration file:
.reliant.jsonin application data directory - Encrypted with system keychain integration (macOS Keychain, Windows Credential Manager, Linux Secret Service)
Key Security Best Practices
Do:
- ✅ Use environment variables for CI/CD
- ✅ Rotate keys regularly (every 90 days)
- ✅ Use different keys for dev and production
- ✅ Monitor API key usage in provider dashboards
Don’t:
- ❌ Commit API keys to git repositories
- ❌ Share configuration files containing keys
- ❌ Use production keys in development
- ❌ Store keys in plain text files
Revoking Compromised Keys
If an API key is compromised:
- Immediately revoke the key in your provider’s dashboard
- Generate a new key
- Update Reliant configuration with the new key
- Review recent activity in provider’s usage logs
- Audit git history to ensure key wasn’t committed
Enterprise Security
For enterprise deployments, Reliant supports additional security features:
Single Sign-On (SSO)
Integrate with your company’s identity provider:
- SAML 2.0
- OAuth 2.0 / OpenID Connect
- LDAP / Active Directory
Contact enterprise@reliant.ai for setup assistance.
Audit Logging
Enable comprehensive audit logging:
{
\"audit\": {
\"enabled\": true,
\"logPath\": \"/var/log/reliant/audit.log\",
\"includeToolCalls\": true,
\"includeFileAccess\": true
}
}
Audit logs include:
- All tool executions
- File access patterns
- API calls made
- Configuration changes
- User actions
Network Policies
Restrict network access:
{
\"network\": {
\"allowedDomains\": [
\"api.anthropic.com\",
\"api.openai.com\"
],
\"blockAllOthers\": true,
\"requireProxy\": true,
\"proxyUrl\": \"http://corporate-proxy:8080\"
}
}
Data Residency
For compliance with regional data laws:
{
\"dataResidency\": {
\"region\": \"eu-west-1\",
\"enforceRegion\": true
}
}
Ensures all data processing and storage occurs in specified region.
Compliance
Reliant is designed to help meet compliance requirements:
GDPR (General Data Protection Regulation)
- Right to Access: Export all your data via Settings → Privacy → Export Data
- Right to Deletion: Delete all data via Settings → Privacy → Delete All Data
- Right to Portability: Data is stored in standard formats (JSON, SQLite)
- Privacy by Design: Minimal data collection, local-first architecture
SOC 2
Enterprise deployments can be configured for SOC 2 compliance:
- Audit logging
- Access controls
- Encryption at rest and in transit
- Regular security assessments
HIPAA
For healthcare organizations:
- Enable security mode
- Use offline mode or HIPAA-compliant AI providers
- Configure audit logging
- Implement data retention policies
Contact enterprise@reliant.ai for compliance assistance.
Encryption
Data at Rest
Local Database:
- SQLite database encrypted with SQLCipher
- AES-256 encryption
- Per-user encryption keys
Configuration Files:
- Sensitive fields encrypted with system keychain
- Platform-specific secure storage integration
Data in Transit
AI Provider Communication:
- TLS 1.3 for all API communications
- Certificate pinning for known providers
- Secure websocket connections (WSS)
Remote Sync (if enabled):
- End-to-end encryption
- Keys never leave your device
- Zero-knowledge architecture
Security Updates
Update Policy
Reliant releases security updates promptly:
- Critical vulnerabilities: Patch within 24 hours
- High severity: Patch within 7 days
- Medium severity: Patch within 30 days
Automatic Updates
Enable automatic security updates:
{
\"updates\": {
\"automatic\": true,
\"securityOnly\": true
}
}
Security Notifications
Subscribe to security notifications:
- Email: security@reliant.ai
- GitHub Security Advisories
- RSS feed: https://reliant.ai/security.rss
Reporting Security Issues
Found a security vulnerability?
Please report it responsibly:
Email: security@reliant.ai
Include:
- Vulnerability description
- Steps to reproduce
- Potential impact
- Your contact information (optional)
Response Time:
- Initial response within 24 hours
- Assessment within 3 business days
- Resolution timeline based on severity
Bug Bounty: We reward security researchers who responsibly disclose vulnerabilities. Contact security@reliant.ai for bounty details.
Best Practices Summary
For Individual Developers:
- Keep API keys secure and rotate regularly
- Enable telemetry opt-out if desired
- Regularly update Reliant to latest version
- Use security mode for sensitive projects
- Review file access permissions for agents
For Teams:
- Use separate API keys per team member
- Enable audit logging
- Implement access controls
- Regular security training
- Establish data retention policies
For Enterprises:
- Deploy with SSO integration
- Configure network restrictions
- Enable comprehensive audit logging
- Implement data residency requirements
- Regular security assessments
Related Topics
- API Keys - Secure API key management
- MCP Servers - Security considerations for external tools
- Troubleshooting - Debug privacy and security issues