Security & Trust
Built for enterprise. Designed to protect.
NeuBird queries telemetry in place and never copies it, so production data never leaves your environment. Here is how it works.
Foundation
The Two Golden Rules
Your data never reaches an LLM
The LLM never sees your specific telemetry with sensitive data such as IP addresses and personally identifiable information. Sensitive identifiers are stripped before any context is shared. The LLM advises on strategy; GenDB executes queries inside your VPC.
No IP leaves your environment
We never ask an LLM to generate code or logic that becomes part of your product. NeuBird uses LLMs for reasoning and pattern recognition only; your intellectual property stays yours.
How It Works
Secure by architecture, not by policy.
GenDB is the data layer that enforces NeuBird's security model at every step. It separates the reasoning layer from the execution layer, so the LLM never sees sensitive identifiers like IP addresses or PII, and your data never leaves your network.
Your Telemetry
Metrics, logs, traces and events stay entirely within your VPC. GenDB reads them locally.
Hypothetical Query
GenDB strips sensitive identifiers (IP addresses, PII, and other private fields) before sending operational context to the LLM.
Local Execution
The LLM returns a strategy. GenDB executes it against your real data locally. Nothing crosses the boundary.
Enterprise Grade
Security controls at every layer.
VPC Isolation
NeuBird deploys within your cloud environment. Data analysis happens on your infrastructure, not ours.
Role-Based Access Control
Granular permissions ensure agents only access the systems and actions explicitly granted to them.
Full Audit Trail
Every observation, inference, and action is logged with timestamps, context, and outcomes. Full records are available for compliance review.
Approval Gates
High-risk actions require human confirmation before execution. You define what needs approval.
Blast Radius Limits
Agents operate within defined boundaries. Configurable limits prevent any single action from affecting more than a specified scope.
Encrypted in Transit
All communication between NeuBird components uses TLS. Credentials are never stored in plaintext.
Responsible AI
Autonomous does not mean unchecked.
Every agent in the center, NeuBird's own investigations and the ones you connect, runs on earned write access. It starts by suggesting, stages changes for approval, and acts only within a scope it has earned, with a complete record of everything it does. The goal is augmentation, not replacement. Your team stays in control of decisions that matter.
- Human-in-the-loop for critical actions
- Configurable autonomy levels per environment
- Automatic rollback on metric degradation
- No lateral movement outside defined scope
- Transparent reasoning at every step
Questions about security?
Talk to our team about your specific compliance requirements, deployment architecture, or enterprise security needs.