BEEM MCP Server — Setup Guide
Connect the BEEM data platform to AI assistants. Explore workspaces, query datasets, build SQL views, monitor pipelines, and manage dashboards — all through natural language.
Server URL
Use this URL to connect any MCP-compatible client:
<https://mcp.beemdata.com/mcp>
Setup
Claude.ai (Web)
- Go to Settings → Integrations
- Click Add MCP Server
- Paste the server URL above
- Sign in with your BEEM account
Claude for Teams & Enterprise (Organization-wide)
On Team and Enterprise plans, an Owner or Primary Owner must add the connector at the organization level before members can use it. This makes BEEM available to the whole team without requiring individual setup.
Step 1 — Owner: Add the connector to your organization
- Sign in to Claude and go to Organization settings → Connectors
- Click Add
- Paste the server URL:
https://mcp.beemdata.com/mcp - Optionally click Advanced settings to provide a custom OAuth Client ID and Secret
- Click Add to confirm
The connector now appears for all members of your organization with a Custom label.
Step 2 — Members: Connect your individual account
Each team member authenticates independently to ensure Claude only accesses data their BEEM account has permission to see.
- In Claude, click the + button at the bottom left of the chat interface
- Select Connectors
- Find the BEEM connector (labelled Custom) in the list
- Click Connect and sign in with your BEEM credentials
- Enable the connector per conversation using the same toggle
Note: Owners can manage connector access and disable specific tool calls under Organization settings → Connectors.
Claude Desktop
- Open Settings → Developer → Edit Config
- Add the following JSON configuration:
{
"mcpServers": {
"beemdata": {
"url": "<https://mcp.beemdata.com/mcp>"
}
}
}
- Restart Claude Desktop
- Authenticate on first use
Claude Code (CLI)
Run this command in your terminal:
claude mcp add --transport http beemdata <https://mcp.beemdata.com/mcp>
A browser window will open to authenticate. BEEM tools are available immediately after.
Other MCP Clients
Any Streamable HTTP MCP client works:
- Use the server URL above
- Auth: OAuth 2.0 with PKCE
- Dynamic client registration supported
- OAuth discovery endpoint:
https://mcp.beemdata.com/.well-known/oauth-authorization-server
Examples
Explore and query data
"Show me the datasets in my workspace and query the first one"
- Retrieves your workspaces and lists available datasets
- Fetches dataset details including SQL query and schema
- Queries dataset content and returns sample rows
Create and deploy a SQL dataset
"Create a dataset called 'Monthly Revenue' that aggregates revenue by month, then deploy it"
- Creates a new dataset in your workspace
- Saves a SQL query as draft and previews results
- Deploys the query as a live Redshift view with scheduling
Monitor pipeline health
"Check if any of my data pipelines failed in the last 24 hours"
- Lists recent pipeline events filtered by failure status
- Identifies affected datasets and data sources
- Shows error details and current DPU usage
Available Tools
Workspaces & Organizations
| Tool | Description |
|---|---|
beem_get_current_user |
Get your user info, workspaces, and organizations |
beem_get_workspace |
Get workspace details including users |
beem_list_workspaces |
List all workspaces in an organization |
beem_get_organization |
Get organization details |
beem_list_user_organizations |
List your organizations |
Datasets
| Tool | Description |
|---|---|
beem_get_dataset |
Get dataset details, metadata, and configuration |
beem_list_datasets |
List datasets in a workspace or folder |
beem_create_dataset |
Create a new dataset (starts as draft) |
beem_update_dataset |
Update description, folder, or AI search settings |
beem_rename_dataset |
Rename and redeploy the Redshift view |
Dataset SQL & Content
| Tool | Description |
|---|---|
beem_get_dataset_sql_queries |
Get SQL query history (deployed and draft) |
beem_save_dataset_sql_query |
Save a SQL query as draft |
beem_deploy_dataset_sql_query |
Deploy SQL — creates Redshift view and scheduling |
beem_get_dataset_content |
Query and retrieve rows from a dataset |
beem_preview_sql_query |
Preview SQL results without deploying |
beem_refresh_dataset |
Trigger a manual dataset refresh |
Dataset Tests
| Tool | Description |
|---|---|
beem_get_dataset_test |
Get test details |
beem_list_dataset_tests |
List all tests for a dataset |
beem_create_dataset_test |
Create a data quality test with SQL |
beem_update_dataset_test |
Rename a test |
beem_delete_dataset_test |
Delete a test and its query history |
beem_save_dataset_test_sql_query |
Save test SQL as draft |
beem_deploy_dataset_test_sql_query |
Deploy a test SQL query |
Sources & Ingestion
| Tool | Description |
|---|---|
beem_get_source |
Get source (catalog) connection details |
beem_list_sources |
List all sources in a workspace |
beem_get_data_object |
Get source table details |
beem_list_data_objects |
List tables for a source |
beem_create_data_object |
Create a new source table |
beem_start_ingestion |
Start catalog sync for a data object |
Dashboards & Visualizations
| Tool | Description |
|---|---|
beem_get_dashboard |
Get dashboard details |
beem_list_dashboards |
List dashboards in a workspace |
beem_create_dashboard |
Create a new dashboard |
beem_update_dashboard |
Rename a dashboard |
beem_get_dashboard_widget |
Get widget details |
beem_list_dashboard_widgets |
List widgets in a dashboard |
beem_create_dashboard_widget |
Add a widget to a dashboard |
beem_update_dashboard_widget |
Update a widget |
beem_get_visualization |
Get visualization details |
beem_list_visualizations_by_dataset |
List visualizations for a dataset |
beem_list_visualizations_by_workspace |
List all visualizations in a workspace |
beem_create_visualization |
Create a visualization |
beem_update_visualization |
Update a visualization |
beem_delete_visualization |
Delete a visualization |
Folders
| Tool | Description |
|---|---|
beem_get_folder |
Get folder details with subfolders and datasets |
beem_list_folders |
List folders within a parent folder |
beem_create_folder |
Create a new folder |
beem_update_folder |
Rename or move a folder |
Flows
| Tool | Description |
|---|---|
beem_get_flow |
Get flow details (dataset-to-destination connection) |
beem_list_flows |
List all flows in a workspace |
Events & Usage
| Tool | Description |
|---|---|
beem_list_events_by_workspace |
List pipeline events in a workspace |
beem_list_events_by_object |
List events for a specific dataset or source |
beem_list_events_by_user |
List events triggered by a user |
beem_get_dpu_usage |
Get DPU consumption and billing cycle info |
Troubleshooting
"Session terminated" or "Session expired"
Your session may have been lost after a server update. Simply retry — the client will re-authenticate automatically.
Authentication fails or loops
Make sure you're signing in with a BEEM account that has access to at least one workspace. Try disconnecting and reconnecting the MCP server.
Tool returns "not found" errors
Verify you have access to the workspace or dataset. The MCP server uses your BEEM permissions — you can only access resources your account has access to.
Need help? Contact team@beemdata.com