Security isn't a feature.
It's the foundation.
Klairr is the AI Analyst for teams that take data seriously. Every layer — from how we query your data to how we store credentials — is designed around a simple principle: your data stays yours.
Don't take our word for it. Ask your AI to summarise this page.
We're an AI Analyst — we know how you do research now. Open this page in the assistant of your choice and ask it to read, summarise, or stress-test our security posture in plain English.
Core principles
Your data stays yours
We never copy your database. Your full datasets remain in your own infrastructure.
We don't replicate your databases
Klairr connects to your data sources with read-only credentials and runs queries on demand. We don't copy, replicate, or migrate your full datasets — only the question, the generated query, and a small result sample needed to display the answer are stored.
Write operations are impossible
INSERT, UPDATE, DELETE, and DROP are blocked at the application layer before reaching your database. We recommend read-only credentials as a second line of defense.
Minimal data retention
Only question text, metadata, and a small sample of query results are persisted. Full datasets remain in your own systems at all times.
No AI training on your data
Our AI provider's terms explicitly prohibit using customer inputs or outputs for model training. Your business data is never used to improve any AI model.
Encrypted end to end
Strong TLS in transit. Industry-standard encryption for credentials at rest. All application data encrypted with tenant-level isolation.
SOC 2 ready
Audit logging, role-based access controls, and encryption are in place today. Our controls and policies are mapped to SOC 2 Trust Services Criteria; we have not yet completed a Type I or Type II audit.
Tenant isolation
Every organization is fully isolated. Separate credentials, scoped data access, and independent AI Memory per tenant.
GDPR compliant
Data-processing practices compliant with GDPR. Article 28 Data Processing Agreement available. Data subject rights supported.
Access control
Granular permissions by role
Four roles with scoped permissions. Every user sees only what they need.
- Full admin access
- User & role management
- Connector configuration
- GRC audit dashboard
- AI Memory management
- Spend limit controls
- Ask questions across connectors
- Query editor
- Create & share reports
- Configure AI Memory
- View usage analytics
- Ask questions
- Save & share reports
- Query editor
- Access assigned connectors
- Ask questions
- View shared reports
- Access assigned connectors
Governance
Full visibility. Complete control.
Built-in governance gives admins a real-time view of every question asked, every action taken, and every risk signal detected.
Complete audit trail
Every question, query, data source, user, and timestamp is logged and searchable. No action goes unrecorded.
Admin action logging
User management changes, connector modifications, AI Memory edits, and role assignments are tracked with full attribution.
Risk signal detection
Automated monitoring surfaces anomalies — unusual query patterns, low-confidence streaks, and access irregularities.
Compliance export
Export audit logs as CSV for external compliance tools, SOC reviews, or internal reporting. Filter by date range for targeted investigations.
Connector-level permissions
Control which data sources each role can access. Production connectors are flagged with a visual warning to reinforce awareness.
Cost controls
Per-query and daily spend limits enforced by role. Query cost estimation and dry-run mode prevent runaway costs before execution.
AI safety
AI you can verify
Every AI-generated answer is grounded in real data, scored for confidence, and fully transparent. Nothing is fabricated. Nothing is hidden.
Grounded answers only
Every answer is traceable to a real query against your actual data. If the data isn't there, Klairr says so. Fabrication is architecturally prevented.
Confidence scoring
Every answer receives a confidence assessment — High, Check, Low, or Failed — with an explanation. Users see answer quality before acting on it.
Full query transparency
Every answer shows its work. Power users can inspect, edit, and re-run the underlying query. Nothing is hidden behind a black box.
Query guardrails
Generated queries are validated before execution. DML and DDL are blocked. Automatic LIMIT injection and byte budgets prevent excessive data retrieval.
Compliance
Regulatory readiness
Klairr is built to meet the compliance requirements of regulated industries. All legal documents are publicly available.
GDPR
Data processing under GDPR Article 28. Data subject rights (access, deletion, portability) supported. Data Processing Agreement available.
Data Processing Agreement
Standard DPA available for all customers. Covers data processing terms, sub-processor disclosure, and breach notification procedures.
View DPA →Privacy Policy
Transparent documentation of what data we collect, why, and how long we retain it. Updated regularly as our practices evolve.
View Privacy Policy →Terms of Service
Clear terms governing product use, data ownership, liability, and SLA commitments.
View Terms →FAQ
Common security questions
Does Klairr store my company data?
Does the AI train on my data?
Can Klairr modify my database?
Is Klairr GDPR compliant?
Is Klairr SOC 2 certified?
Where can I find your legal documents?
Who do I contact about security concerns?
Ready to evaluate Klairr for your organization?
Our team is available to discuss security architecture, compliance requirements, and deployment options in detail.
No credit card required