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

# Live Dashboard overview

> Generate business dashboards from natural language, refine them in chat, filter and refresh in Workbench, and share formal versions with your team.

Live Dashboard, also called Artifacts Dashboard, turns a business request into a saved dashboard that you can refine and refresh. Choose standard charts or ask for an HTML page with a custom layout, colors and interactions. The dashboard remains one Artifact as you edit its draft, publish formal versions and share it.

For example, request quarterly net sales by month and region, with a total at the top, a trend on the left, details on the right and a region filter. The Assistant checks available data and metric definitions before creating a draft.

<CardGroup cols={2}>
  <Card title="Create your first dashboard" icon="chart-column" href="/en/ai/live-dashboard/create">
    Prepare data and business rules, describe a page in natural language and check the result in Workbench.
  </Card>

  <Card title="Share with your team" icon="users" href="/en/ai/live-dashboard/sharing">
    Publish a reviewed version so members can filter and refresh authorized data on a standalone page.
  </Card>
</CardGroup>

## When to use it

* **Business and metric tracking:** combine totals, trends and details on one page, with filters such as region or period.
* **File analysis summaries:** clean and aggregate authorized workspace files using explicit metric definitions, then visualize the results.
* **Ongoing review and team viewing:** refine the same dashboard's presentation or analysis requirements, then share a human-confirmed formal version.

## A dashboard's workflow

1. **Create:** choose an Assistant with dashboard capabilities and describe the source, metric, period and layout.
2. **View:** after a successful save, click **Open** on the chat resource card to open **Live dashboards** in ChatKit's Workbench.
3. **Edit:** open the target and continue describing changes. They are saved as a draft under the same dashboard identity.
4. **Publish:** check the data and interactions, select **Publish draft**, then confirm a formal version.
5. **Share:** enable read-only organization sharing. Members can filter and refresh data they are authorized to access.

The chat resource card provides a return point: **Open** takes you directly into ChatKit's own Workbench. A team sharing link opens a standalone dashboard. Its toolbar collapses into a translucent button and expands on hover or click to leave more room for viewing.

## What you can create

| Mode | Suitable uses |
| - | - |
| Standard charts | Bar, line, area, pie and scatter charts, plus data tables |
| Custom page | Business summaries, KPI layouts, trends and details with ECharts |
| File or script results | Python/JavaScript processing of authorized files or sources |

Custom pages control their layout but use configured data access. They cannot connect to arbitrary websites, load CDN scripts or read platform credentials. Three.js core is available; 3D pages still need separate browser and GPU compatibility checks.

## What live data and versions mean

* Opening, filtering and refreshing query data again. Fetch time is not necessarily the last business-data modification time.
* Custom HTML can include a refresh timer. There is no universal automatic-refresh switch or guaranteed change-stream push subscription.
* Drafts and formal versions save the page definition, metric configuration and library dependencies, not all query results at that moment.
* Historical versions query current accessible data with their older definition. They are not historical data snapshots.

Team sharing is read-only and requires a formal version. Drafts remain owner-managed. Workspace script dashboards currently cannot be shared with a team.

## Start here

* [Create a dashboard from a request](./create)
* [Edit and publish](./edit)
* [Data sources, files and refresh](./data-sources)
* [Team sharing and standalone viewing](./sharing)
* [Troubleshooting](./troubleshooting)

If sources or dashboard tools are missing, ask an administrator to configure the required plugins, permissions and runtimes. Availability depends on your deployment.

## Related capabilities

* [Assistant conversations](/en/ai/agent/conversation/conversation) capture requests and ongoing edits; Live Dashboard saves, displays and refreshes the resulting analysis page.
* [Artifacts](/en/ai/agent/artifacts/index) supplies result identity, drafts, versions and access management. This section covers the dashboard-specific workflow.
* [Ontology Workbench and resource chat](/en/data/features/agent-execution/workbench-and-chatkit) explains data-xpert resource context. Dashboards can use configured ontology sources, subject to their permissions and permitted read-only operations.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.