The Connected Architecture
How the Four Products Share
Identity, Data & Context.
The four products share single sign-on, roles, OpenTelemetry data, and MCP tools, so a finding in one product reaches the others.
One Finding, Four Products
The products communicate through standardized OpenTelemetry streams, scoped JWT contexts, and Model Context Protocol interfaces. A vulnerability detected in TrustIvan triggers an evidence update in trustlynce, logs an event in aiAxonIQ, and alerts MITHUNAI agents.
One sign-on, one set of roles, and shared context across every product.
What Each Product Contributes
Detailed specifications of how each layer operates and protects enterprise workloads.
The Telemetry & Observability Plane
Receives OpenTelemetry logs, metrics, and traces over OTLP, buffers them in Kafka, stamps each record with a tenant ID, and writes them to partitioned ClickHouse tables for fast analytical queries.
The Continuous GRC & Evidence Plane
Turns cloud state, scan results, and incident records into verifiable, audit-ready evidence. Its 270+ Model Context Protocol (MCP) tools let LLM agents query controls, run risk assessments, and draft SOC 2 evidence packs for review.
The AppSec & Machine-Identity Plane
Secures code, containers, and non-human identities. Scans container images and SBOMs with Trivy, and checks every tool call from AI agents and MCP servers against policy at runtime to limit lateral movement.
The Enterprise AI Plane
Runs private model routing, document ingestion, and retrieval with citations. Answers link to the internal sources they came from, and agents act only through the tools and scopes the other planes allow.
Identity: SAML 2.0, OIDC & Scoped JWTs
When an engineer signs in through Okta, Microsoft Entra ID, or Google, each product receives a short-lived, signed token with that person's roles. TrustIvan applies AppSec permissions, trustlynce assigns auditor roles, aiAxonIQ restricts queries to the tenant's ClickHouse partitions, and MITHUNAI scopes AI workspaces, with no duplicate accounts.
No Shared Secrets
Products never share signing keys. Services verify identity through signed internal endpoints.
Tenant-Level Data Isolation
ClickHouse, PostgreSQL, and vector stores keep each tenant's data in separate partitions with their own encryption keys.
Scoped MCP Tool Access
AI agents are granted ephemeral, time-bounded tokens limited strictly to the tools and data required for their task.
Talk to an Architect
See the Four Products Running in Your Environment.
A technical session with a Lionhouse architect covering your stack, your deployment model, and a pilot plan.