Who they are
Engineers, not consultants
The people who build Lustena, working inside your rollout. They write the policies, connect the SDK and fix what breaks, with commit access to the product.
Enterprise
The Enterprise plan,
with forward deployed engineers
The full Lustena platform, rolled out by the engineers who build it. In weeks, agents are found, owners named, limits set, findings routed and evidence ready for your auditor.
Onboarding
How the platform goes live
Four phases, from the first device to the evidence package.
01
Roll out
The browser extension goes out through your MDM, devices enroll by team, and the agents your engineers already run are connected.
02
Map
Lustena builds the inventory as agents are used: every agent, its owner, its connectors, and what is still outside coverage.
03
Govern
The first limits per agent are set with the people who own the risk, and alerts go to the channels you already watch.
04
Run
Your team runs the platform: limits evaluated on every session, findings routed to owners, and an evidence package for the framework you chose.
What you get
A control plane that runs on day one
Every agent found and assigned an owner, limits evaluated on every session, findings routed to owners, and evidence prepared for GDPR, the EU AI Act, ISO/IEC 42001, NIST AI RMF, AIUC-1, LGPD or your own policies.
How long
Weeks, not quarters
The browser rollout takes an hour. The rest of onboarding covers what most teams can't staff on their own: owners, limits, cost visibility per agent, and the habits of the people who use AI every day.
Why it works
The platform does the governing. Our engineers make sure it runs on day one.
The platform
What stays when the engineers leave
The same control plane every customer runs, operated by your team.
For IT and security
AI governance control plane
See every agent your people use, put an owner on each, apply limits and keep one audit trail. Rolled out by policy, nothing installed inside the agents.
- Any AI app in a managed browser
- Google Admin, Intune or Jamf
- Inventory, limits, findings, evidence
For engineering
Developer surface
Instrument the agents your teams build, register coding assistants and red-team their endpoints. Everything lands in the same inventory, governed the same way.
- Python SDK for CrewAI, LangGraph, ADK, Agno
- MCP server for Claude Code, Cursor, Codex
- Scanner with ten agent-native checks
Security practices
Built to be audited
Lustena does not yet hold certifications of its own and will not print a badge it did not earn. These are the practices an auditor can verify.
Content-free by construction
By default, prompt text, file contents and file names never leave the device. Only categories, counts, destinations, channels and HMAC fingerprints do.
Tenant isolation in code
Each customer's data is isolated in code: every read is checked against the requesting company and user, and a device can only submit its own findings, nothing else.
Encryption and retention
Content capture is off by default. When you turn it on, redacted content is encrypted at rest, kept only for the retention you set, and every read is logged.
Frequently asked questions
What is a forward deployed engineer?
A Lustena engineer who works inside your rollout during onboarding: rolling out the sensors, building the inventory with your teams, writing the first limits and preparing the evidence. They are the same people who build the product.
How long does onboarding take?
Weeks, not quarters. The first phase covers one rollout and one framework. The browser rollout itself takes an hour; the rest covers owners, limits and evidence for the teams in scope.
Which AI apps and agents are covered?
In the browser: ChatGPT, Claude, Gemini and Copilot. In code: CrewAI, LangGraph, Google ADK, Agno and OpenTelemetry sources. In developer tools: Claude Code, Cursor and Codex. On the desktop: the ChatGPT, Claude and Copilot apps.
Does the Enterprise plan certify us?
No. Independent assessors certify. Lustena prepares the evidence for the framework you chose, GDPR, the EU AI Act, ISO/IEC 42001, NIST AI RMF, AIUC-1, LGPD or your own policies, so the assessment is faster.
Can Lustena be self-hosted?
Yes. The control plane runs as a managed service on Google Cloud by default, and on the Enterprise plan it can be deployed in your own cloud or on premises.