Skip to main content

Console tour

The QuilrAI console at web.quilr.ai is organised around what you manage (people, applications, agents, interactions and controls), not around the sensors that collect the data. This page shows how the sidebar is laid out and where each job lives.

How sensors feed every page​

The LLM Gateway, MCP Gateway, Endpoint Agent and Browser Extension, plus connected platforms such as OpenAI Compliance, GitHub, Microsoft Copilot Studio and Azure AI Foundry, all report into the same data. You do not visit one page per sensor. Instead:

  • every Observe page combines all sources, and most lists have a Sensor or Sources filter to narrow them,
  • rows and drawers name the sensor that saw the activity,
  • each sensor keeps one home under Settings (deployment and operation) and one tab under Govern > Policy Engine (enforcement).

To hide a source from the console without stopping ingestion, use Data sources.

The sidebar​

The sidebar has four groups: Observe, Govern, Assessments and Settings.

Observe​

PageWhat it is for
OverviewExecutive snapshot: AI adoption, the agentic estate, and posture and risk.
Costs & SavingsAI spend and tokens by source, who drives them, and settings that would save tokens.
GraphInteractive map of how people, apps, agents, MCP servers and data connect.
Findings & InteractionsEvery finding and every AI interaction, with an investigation drawer. Includes the Triage center.
UsersEach person's AI usage and risk, plus browser and endpoint deployment coverage.
InventoryCatalog of every AI asset Quilr discovered, with approval status and risk.
AgentsEvery AI agent observed, with the skills, MCP servers, prompts and people behind it.
DashboardsCustom dashboards you describe in plain language to an AI designer.

Govern​

PageWhat it is for
Policy EngineWhat happens when AI use crosses a line, per enforcement surface: Browser Extension, LLM Gateway, MCP Gateway and Endpoint Agent.
Detection ModelsWhat counts as sensitive data or an adversarial prompt: built-in and custom detectors.
Action RequestApprove or reject justification requests raised when the browser extension blocked an action.

Assessments​

Red Teaming actively tests LLM applications, MCP servers, agents and models for weaknesses.

Settings​

GroupPages
AI GatewayWorkflow Agents, LLM Gateway, MCP Gateway, Models, Skills Library
OrganizationGeneral, Smart Groups, Roles & Permissions, Single sign-on, Organizational Policies, Data Sources, Audit Logs
Data managementData Retention, Export Center
OtherIntegrations, Endpoint Agent, Browser Extension, User Interaction Hub, User Profile and Display

Search, Quilr Assist and the page guide​

  • Search at the top of the sidebar (Cmd+K) opens a command palette to jump to any page.
  • The Quilr Assist bubble in the bottom-right corner opens the in-console assistant.
  • The ? button next to it opens the page guide for the page you are on.

You can hide the bubble, turn off the page guide, or make the bubble compact under Settings > Display. See Display and profile.

Patterns shared by most pages (period selector, filters, saved views, layouts, drawers and export) are covered in Views, filters and drawers.

Self Service for end users​

Switch to Self Service at the bottom of the sidebar opens the end-user portal. There, people connect their AI apps, see the LLM Gateway applications they are approved for, and use Workflow Agents shared with them. See Self Service.

Next steps​

On this page