1. Enroll
Devices enroll through Google Admin, Intune or Jamf with an enrollment token tied to the employee's corporate email. The token never leaves managed storage.
Connect · Agent Identity
Every agent gets a name,
an owner and a history.
Lustena gives every AI agent your company runs a persistent identity and a named owner, whether engineering built it or a sales rep set it up in ChatGPT in ten minutes. Access and accountability follow the person behind it.
Works with
ChatGPTClaudeGeminiCopilotClaude CodeCursorCrewAILangGraphYou can't govern what has no owner
Lustena puts a name on every agent.
Without a control plane On the control plane
Ownership
Without a control plane: Agents live in personal accounts, shared logins and no-code tools. Nobody can say who is responsible for what an agent did last week.
On the control plane: Each agent has a declared owner and team. Observed agents are attributed to the device and the employee who used them, until someone claims them.
Inventory
Without a control plane: The list of agents is whatever the last survey said. Custom GPTs, projects and connectors appear and disappear without a record.
On the control plane: One inventory of declared and observed agents, with first and last seen, connectors used and sessions per week, kept current by the sensors.
Attribution
Without a control plane: When a customer list leaves in a prompt, the only evidence is a screenshot and a memory.
On the control plane: Every finding and every alert is attributed to a person, an agent and a session, in an exportable audit trail.
Inventory
One inventory of declared and observed agents
Production2026-10-14 · 02:41:07Z
Agents
Every agent this workspace knows about, declared by a person or observed in use, with its owner and connectors.
Inventory · 5 rows
| Agent | Owner · connectors | Status |
|---|---|---|
| GPT · Proposalsana@acme.com | ana@acme.comGmail, Drive | Declared |
| Claude · Finance projectjoao@acme.com | joao@acme.comSheets | Observed |
| Support bot (LangGraph)platform team | platform teamZendesk API | Declared |
| Copilot Studio · HR FAQunknown | unknownSharePoint | Observed |
| Claude Code · repo-assistengineering | engineeringMCP: GitHub | Declared |
Selected · GPT · Proposals
Owner
ana@acme.com · Marketing
Limits
v3 · tools, hosts, data
Last seen
2m ago · chatgpt.com
Identity for agents, from the device up
The identity layer starts where the agent is used and ends in an audit trail your security team can read.
2. Observe
Each AI app, custom GPT, project or connector a person uses becomes an observed agent, keyed by what the browser sees, never by content.
3. Declare
Owners claim agents, add a purpose and a team, and discard what is not an agent. Declared and observed stay distinct in every view.
4. Attribute
Findings, limits and sessions are attributed to the agent and the person. Developers connect code agents with the SDK or the MCP server under the same identity model.
Sensors
- Browser extension by policy
- Desktop collector for AI apps (Enterprise plan)
- Python SDK for frameworks
- MCP server for developer tools
- Scanner for endpoints
Identity model
- Declared and observed agents
- Owner, team and purpose
- Device and employee attribution
- Lineage for runtime agents
Lifecycle
- Claim, declare, discard
- First and last seen
- Connectors and hosts used
- Sessions per week
Access
- Enrollment tokens per team
- Device revocation
- Scoped API tokens
- Passwordless workspace login
Your identity provider stops at the login.
SSO knows who signed in to ChatGPT. It does not know which GPT was created, what was pasted, or which connector was authorized afterwards. Lustena picks up where the login ends.
Outside the loop · where legacy controls sit
- IAM & SSO
- MDM & devices
- SaaS admin consoles
- HR directory
No line of sight inside
Inside the loop · where Lustena sits
Agent
every action →Lustena← result · next step
System
The agent acts here: prompt, paste, upload, connector, tool call.
Questions
Does Lustena replace my identity provider?
No. Devices enroll with an enrollment token and carry the employee's corporate email. Agents get their own identity on top of the people who use them. On the Enterprise plan, platform login goes through your identity provider with SSO.
What is an observed agent?
An agent Lustena saw in use before anyone declared it: a custom GPT, a project, a connector, or an actor emitting events through the SDK. Observed agents are listed separately until an owner claims or discards them.
Can an employee see what is recorded about them?
Yes. The extension's own findings page shows every finding recorded on that device, whether the device is in local mode or enrolled, and whether content capture is on. Admins can exclude apps and sites from observation, and an employee can ask for their records to be deleted.