> ## Documentation Index
> Fetch the complete documentation index at: https://docs.originhq.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Canvases

> Multi-kernel dashboards — compose, arrange, and share custom analytics views.

<Frame>
  <img src="https://mintcdn.com/originhq/BoZcxy2aZcnPe0ai/images/origin-canvases-1.png?fit=max&auto=format&n=BoZcxy2aZcnPe0ai&q=85&s=c99f700d9848f6a9217d48f41f5d5e97" alt="Canvases" width="1440" height="810" data-path="images/origin-canvases-1.png" />
</Frame>

**Canvases** are custom analytics dashboards built from one or more Kernels. A Canvas lets you compose exactly the view you need — whether that's an executive summary of AI spend, a security incident dashboard, or a per-team adoption report — and save it for reuse or sharing.

## What a Canvas contains

Each Canvas is a collection of **Kernels** arranged in a grid layout. Each Kernel is a self-contained data visualization — a bar chart, a table, a time series — that runs a specific analytic query against your Origin data.

Canvases are:

* **Scoped** — they inherit whatever scope (agent, endpoint group, date range) is selected in the left panel
* **Live** — every time you open a Canvas it reruns all its Kernels against fresh data
* **Shareable** — you can share a Canvas URL with anyone who has access to your Origin tenant

## Step 1: Navigate to Canvases

Click **CANVASES** in the Views section of the left panel. The center panel switches to the Canvases view.

## Step 2: Review headline metrics

The **AI Usage Overview** canvas is the built-in default canvas. It shows headline counts at the top for the current 30-day window:

* **Prompts** — total prompt count
* **Endpoints** — unique endpoints with activity
* **Sessions** — distinct session count
* **Models** — distinct model identifiers seen

Each metric is sourced from the `/activity-summary` kernel and refreshes on demand.

## Step 3: Explore activity trends

Below the headline metrics, the **Activity trend — prompts per day** section shows a bar chart of daily prompt volume. Hover over any bar to see the exact date and count. Use this to identify:

* Unusual spikes in AI usage
* Weekly patterns (e.g., lower usage on weekends)
* Ramp-up curves after a new tool deployment

## Step 4: Switch between canvas tabs

Use the tab bar at the top of the canvas to switch between views:

* **AI usage overview** — headline counts and trend charts
* **Adoption & concentration** — per-team or per-agent adoption rates and usage concentration metrics

## Step 5: Build a custom canvas

Click the **+** button in the canvas tab bar to create a new Canvas. Give it a name, then use the Kernel browser to add visualizations. You can:

* Add built-in kernels from the Library
* Fork an existing kernel to customize its query
* Arrange kernels by drag-and-drop

See the [Library walkthrough](/walkthroughs/library) for more on how Kernels work.

<Tip>
  The **REFRESH** button next to each kernel's data source indicator reruns that kernel immediately. This is useful when you've just made a configuration change and want to see the updated data without waiting for the next scheduled refresh.
</Tip>
