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

# Reports and Dashboards: Current Availability

> Create Salesforce reports and dashboards through ApexGenius MCP metadata tools, or use the separate web Report Builder available by request.

The current self-serve ApexGenius MCP connector can create and update Salesforce reports and dashboards through its generic metadata path. Your AI client stages native SFDX report or dashboard files, reads them back for review, deploys the selected files through the Salesforce Metadata API, and checks the deployment result.

This is a metadata workflow rather than a dedicated `create_report` or `create_dashboard` tool. The separate ApexGenius web workspace and Report Builder are still available by request, but they are not required for metadata-based creation through MCP.

## What you can do through MCP today

* Query Salesforce records with task-specific SOQL.
* Inspect objects, fields, and relationships.
* Search, read, stage, and deploy metadata, including reports and dashboards.
* Create, update, or delete records when the tool is available and the authorized Salesforce user has permission.
* Review content-minimized operational and write activity.

## Create a report or dashboard through MCP

Use `salesforce_write_file` or `salesforce_edit_file` with the correct SFDX path, such as `reports/MyFolder/Pipeline.report-meta.xml` or a file beneath `dashboards/`. Read the staged file back with `salesforce_read_file`, confirm its folder, filters, groupings, and source report references, then call `salesforce_deploy` with only the intended files. Use `salesforce_check_deploy_status` until Salesforce returns the final component results.

Write confirmation depends on the AI client and its policy controls. Review the risk label and exact staged paths before approving a deployment. ApexGenius records supported deployment activity, but the gateway does not promise a universal confirmation mode across every client.

Set `check_only: true` on `salesforce_deploy` to validate the selected staged files without committing them. Review the returned component and test results, correct any failures, then call the same operation with `check_only: false` when you are ready to deploy. There is no separate validation tool.

## If you prefer a guided report workspace

Use **Request workspace access** for the separate web product. The Report Builder supports a guided report-building workflow, while the MCP route is the direct metadata path from Claude or ChatGPT.

<Columns cols={2}>
  <Card title="Set up the MCP connection" href="/guides/mcp-setup">
    Hosted OAuth setup for Claude, ChatGPT, VS Code, Cursor, and compatible clients.
  </Card>

  <Card title="Full tool reference" href="/reference/mcp-tools">
    The complete current public tool contract.
  </Card>
</Columns>