We manage AI agents. You own the environment they run in. That separation is by design — and it's non-negotiable.
Every agent runs inside your cloud accounts or on-premises servers. We deploy to your environment — we never ask you to send data to ours.
Agent processing happens locally. Inputs, outputs, logs, and intermediate states all remain within your network boundary.
Our management access is limited to agent configuration and performance monitoring. Every action is logged. Access can be revoked at any time.
We provide full deployment manifests, network diagrams, and access logs. If you want to audit us, we'll hand you the keys.
Agents operate within your VPC or private network. No inbound or outbound connections to AfrexAI infrastructure are required for agent operation.
Each customer deployment is completely independent. There is no shared infrastructure, no shared databases, and no cross-tenant access of any kind.
All agent data — configuration, logs, cached outputs — is encrypted using your cloud provider's native encryption (AWS KMS, Azure Key Vault, GCP KMS) or your own key management system.
All communication between agents and integrated systems uses TLS 1.3. The management tunnel between your environment and our monitoring dashboard uses mutual TLS with certificate pinning.
You control all encryption keys. We never have access to your master keys. Agent credentials are stored in your secrets manager — not ours.
Our management access follows the principle of least privilege. We can update agent configurations and view performance metrics — nothing more.
Management access requires SSO integration with your identity provider, multi-factor authentication, and IP allowlisting. We work within your existing IAM policies.
We are actively pursuing SOC 2 Type II certification. Our target completion date is Q3 2026. In the meantime, we provide detailed security questionnaire responses and are happy to work with your security team directly.
Our deployment practices are designed to be compatible with your existing compliance requirements:
We sign DPAs that reflect the reality of our architecture: we manage agents, we don't process your data. Our legal framework is built around this distinction.
We're happy to walk through our architecture with your security and compliance teams.
Book a Security Review