Navigate the Checkpoint dashboard to manage detection, policies, delegations, and deployments

Getting Started

The Checkpoint dashboard is your central hub for managing AI agent detection, enforcement, and governance. After signing in you land on your organization's Home page, and the left rail links to the main surfaces.

Managing Projects

Create, configure, and manage Checkpoint projects for AI agent detection.

Analytics & Reporting

Understand detection analytics, traffic patterns, and AI agent activity.

Project Settings

Configure detection, enforcement, and governance settings per project.

The left rail starts with Home, then groups the product surfaces by pillar under Detect, Enforce, and Govern. Settings and Installations sit at the bottom of the rail, above your account menu.

Home

Item	What it shows
Home	Your organization's landing page: Today's Briefing, a summary of the agents, bots, and pages seen over the last seven days, above a chat with the dashboard assistant

Detect

Item	What it shows
Activity	Each classified request and the decision applied to it, filterable by agent, class, verdict, path, policy, and time, with an inspector for every event
Analytics	Traffic through four lenses (AI, Humans, Bots, Blocked) over 7, 30, or 90 days. Select an agent or bot for its detail page
Review queue	Identity merges escalated for a person to approve or reject (Merge review), and the detections corrected with Mark verdict wrong in Activity (Verdict corrections). Recording a decision doesn't merge or re-score anything by itself

Enforce

Item	What it does
Policy	Author, dry-run, and deploy Cedar policies; configure the authentication and consent experience under Policy → Auth
Delegations	View and revoke the delegations agents hold

Govern

Item	What it does
Agents	The managed agents your organization has hired, which of them are live, and a detail page for each. See Molti
Servers	Launch an MCP server or authorization host from a template, then manage the servers you've deployed. See Deploying a KYA-OS Server
Connections	Your organization's MCP gateway: the upstream MCP servers your agents connect to, their tools, and which members can use them
Audit	The project's signed KYA-OS protocol audit ledger, in the Audit Explorer: recorded protocol events, their verification checks, and signed evidence bundles you can download and verify offline

Settings and Installations

Item	What it does
Settings	Organization settings. The rail switches to a settings menu: Organization, Team, Registry identity (the organization's `did:webvh` in the KnowThat.ai registry), and Single sign-on
Installations	What's set up on the project and what isn't yet, starting with the Detect install flow, whose snippets carry your Project ID and API key (see Credentials)

Your account menu, at the foot of the rail, opens your profile, API keys, integrations, the project list, and billing.

Review queue, Policy, Delegations, Audit, and Installations always open one project: the one selected in the project switcher or, with no project selected (Overview), your organization's primary project. Activity works the same way. Analytics shows the selected project, or the whole organization when no project is selected.

Organizations still on the earlier dashboard see a sidebar instead, with Monitor, Analyze, Agents, and Settings plus a few admin-only items, and land on Monitor after signing in.

Key Areas

Activity

The Activity feed shows classified requests in near real time. Each entry includes:

Timestamp
Detection class (human, ai_agent, bot, incomplete_data)
Confidence score (0–100)
Agent name (if identified)
The policy decision that was applied (or, in observe mode, what it would have been)

Use Activity to verify your integration is working and to investigate specific visitors.

Analytics

Analytics provides aggregated views:

Detection rate: Percentage of traffic classified as agents or bots
Agent type breakdown: Distribution across detection classes
Traffic trends: Historical detection patterns over time
Top agents: Most frequently detected agent types

See Analytics & Reporting for details.

Policy

Policy is where enforcement lives. Describe a rule in plain language in Compose, review the generated Cedar, dry-run it against real traffic, then Authorize & deploy. The Policy → Auth surface configures the authentication and consent journey agents and users see when a policy challenges them. See Policies for the full authoring guide.

Agents & Delegations

Agents lists your organization's own managed agents, with a detail page for each. The agents that visit your properties appear in Analytics and Activity instead: select one in Analytics for its detail page. Delegations shows the authorization grants those agents hold (what scopes, for how long, and for which user) and lets you revoke them. See Govern for the underlying model.

Creating Your First Project

Sign in to the Checkpoint dashboard
Create or select an organization
Create a new project for the domain you want to protect
Follow the guided setup to install a detection method (Pixel, Beacon, Middleware, or Gateway)
Find your credentials and verify the integration under Installations

Each project gets a unique Project ID and API Key. Use these credentials when configuring Pixel, Beacon, Middleware, or Gateway integrations.

See Managing Projects for a detailed walkthrough.

Next Steps

Managing Projects: Create and configure projects
Project Settings: Legacy per-project settings pages (Detect, Govern, Enforce)
Analytics & Reporting: Understand your detection data
Detect: Set up detection methods
Enforce: Configure enforcement policies
Govern: Deploy KYA-OS governance

Dashboard Overview