From omni-analytics
Programmatically create, update, and manage Omni Analytics documents and dashboards via REST API, including lifecycle, tiles, visualizations, filters, layouts, and workbook models.
npx claudepluginhub exploreomni/omni-claude-skillsThis skill uses the workspace's default tool permissions.
Create, update, and manage Omni documents and dashboards programmatically via the REST API — document lifecycle, workbook models, filters, and dashboard content.
Create, update, and manage Omni Analytics documents and dashboards programmatically — document lifecycle, tiles, visualizations, filters, and layouts — using the Omni CLI. Use this skill whenever someone wants to build a dashboard, create a workbook, add tiles or charts, configure dashboard filters, update an existing dashboard's model, set up a KPI view, create visualizations, lay out a dashboard, create a document, rename a workbook, delete a dashboard, move a document to a folder, duplicate a dashboard, or any variant of "build a dashboard for", "create a report showing", "add a chart to", "make a dashboard", "update the dashboard layout", "rename this document", "move to folder", or "delete this dashboard". Also use when modifying dashboard-level model customizations like workbook-specific joins or fields.
Automates Looker dashboards by adding elements, tiles, and filters via Node.js scripts. Useful for data discovery and business intelligence workflows.
Generates self-contained interactive HTML dashboards with KPI cards, charts, filters, and tables from queries, CSVs, or samples for reports and monitoring.
Share bugs, ideas, or general feedback.
Create, update, and manage Omni documents and dashboards programmatically via the REST API — document lifecycle, workbook models, filters, and dashboard content.
Tip: Use
omni-model-explorerto understand available fields andomni-content-explorerto find existing dashboards to modify or learn from.
POST /api/v1/query/run first, including the filters you intend to use. This catches bad field names, missing joins, and malformed filter expressions before they become broken tiles.config, visType, or prefersChart are misconfigured. Default to chartType: "table" for reliable rendering, and configure chart visualizations in the Omni UI.identifier not id for all document API calls — .id is null for workbook-type documents and will silently fail.pivots array is present (reported Omni bug). If boolean filters aren't applying, remove the pivot and test again.PUT /api/v1/documents/{documentId} replaces the entire document state. Always read the existing document first and modify from there, or you'll lose tiles you didn't include.export OMNI_BASE_URL="https://yourorg.omniapp.co"
export OMNI_API_KEY="your-api-key"
When unsure whether an endpoint or parameter exists, fetch the OpenAPI spec:
curl -L "$OMNI_BASE_URL/openapi.json" \
-H "Authorization: Bearer $OMNI_API_KEY"
Use this to verify endpoints, available parameters, and request/response schemas before making calls.
Omni dashboards are built from documents (workbooks). Each has:
Documents can be created with full query and visualization configurations via queryPresentations. Fine-tuning tile layout is best done in the Omni UI.
curl -L -X POST "$OMNI_BASE_URL/api/v1/documents" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"modelId": "your-model-id",
"name": "Q1 Revenue Report"
}'
Returns the new document's identifier, workbookId, and dashboardId.
Use queryPresentations to create a document pre-populated with query tabs and visualization configurations.
Doc gap: The create-document API docs mention queryPresentations but don't show the complete structure. This section documents the full format.
curl -L -X POST "$OMNI_BASE_URL/api/v1/documents" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"modelId": "your-model-id",
"name": "Q1 Revenue Report",
"queryPresentations": [
{
"name": "Monthly Revenue Trend",
"topicName": "order_items",
"prefersChart": true,
"visType": "basic",
"fields": ["order_items.created_at[month]", "order_items.total_revenue"],
"query": {
"table": "order_items",
"fields": ["order_items.created_at[month]", "order_items.total_revenue"],
"sorts": [{ "column_name": "order_items.created_at[month]", "sort_descending": false }],
"filters": { "order_items.created_at": "this quarter" },
"limit": 100,
"join_paths_from_topic_name": "order_items",
"visConfig": { "chartType": "lineColor" }
},
"config": {}
}
]
}'
Tip: Default to
"config": {}for reliable rendering — Omni will auto-generate chart config. For precise chart styling, build a reference dashboard in the UI and read it back viaGET /api/v1/documents/{documentId}. See references/queryPresentations.md for complete config examples by chart type (KPI, line, bar, area, pie, scatter, etc.).
| Parameter | Notes |
|---|---|
modelId | Use the base shared model UUID, not a branchId. Get this from the List Models API. |
| Field format | table.field_name or table.field_name[week|month|day|quarter|year] for time granularity |
sorts | column_name must match the exact field string (e.g., "order_items.created_at[month]"), with sort_descending boolean |
| Parameter | Required | Description |
|---|---|---|
name | Yes | Tile/tab title |
topicName | Recommended | Topic name for the query — set this whenever querying from a topic. Ensures correct join context in the dashboard. |
prefersChart | Yes | Must be true to render a chart. Without this, Omni always shows the results table regardless of any other vis settings. |
visType | Yes | Visualization renderer: "omni-kpi" for KPI tiles, "basic" for all standard charts (line, bar, area, scatter, pie, etc.). |
fields | Yes | Duplicate of query.fields — must be present at this level too. |
config | Yes | Chart-specific configuration object. Shape varies by chart type — read a reference dashboard to get the exact structure. |
chartType | No | Optional chart subtype at the presentation level (e.g. "barGrouped"). |
description | No | Tile description. |
query | Yes | Query definition (see below). |
The query object within each query presentation uses the same structure as the Query API:
| Parameter | Required | Description |
|---|---|---|
table | Yes | Base view name |
fields | Yes | Array of view.field_name references (supports timeframe brackets like [month]) |
sorts | No | Array of { "column_name": "...", "sort_descending": bool } |
filters | No | Object of { "field_name": "expression" } — supports "last 90 days", "this quarter", ">100", etc. |
limit | No | Row limit (default 1000, max 50000) |
join_paths_from_topic_name | Recommended | Topic name for join resolution — set this alongside topicName on the parent queryPresentation. |
pivots | No | Array of field names to pivot on |
Note:
modelIdis not needed inside the query object — it's inherited from the document's top-levelmodelId.
visConfig belongs inside the query object — not at the queryPresentation level. When passed as a sibling of query, it is silently dropped by the API.
visConfig alone does not control chart rendering. It stores the chart type hint on the query, but the actual rendering is driven by prefersChart, visType, and config at the queryPresentation level.
chartType values:
| chartType | Visualization |
|---|---|
kpi | KPI / single value |
lineColor | Line chart |
barColor | Bar chart |
areaColor | Area chart |
stackedBarColor | Stacked bar chart |
pie | Pie / donut chart |
scatter | Scatter plot |
heatmap | Heatmap |
map | Map visualization |
table | Data table |
The config object at the queryPresentation level defines the actual chart rendering. Its structure varies by chart type — see references/queryPresentations.md for complete config examples by chart type.
The most reliable way to get the correct config for a given chart type is to build the chart in the Omni UI and read it back via GET /api/v1/documents/{documentId}.
The most reliable way to learn config, visType, and field names is to read an existing dashboard document:
Step 1: Find a reference dashboard
curl -L "$OMNI_BASE_URL/api/v1/documents" \
-H "Authorization: Bearer $OMNI_API_KEY"
Step 2: Get its full document
curl -L "$OMNI_BASE_URL/api/v1/documents/{documentId}" \
-H "Authorization: Bearer $OMNI_API_KEY"
Returns the complete queryPresentations array including topicName, visConfig, config, and the full query object for each tile — use this as the source of truth when recreating or templating dashboards.
Tip: Build a reference dashboard in the Omni UI with the chart types and styling you want, then read it via
GET /api/v1/documents/{documentId}to capture the exactqueryPresentationsstructure to use as a template.
These apply when copying queryPresentations from an existing document (for both creating new dashboards and updating existing ones):
model_extension_id from each query object — these reference model extensions scoped to the source document and will cause "Chart unavailable" errors.GET /api/v1/documents/{id} returns all queries including workbook-only tabs not shown on the dashboard. Only include the queryPresentations you want as visible tiles.topicName are valid — SQL-mode and tab-selector queries won't have a topicName. Do not add one.curl -L -X PATCH "$OMNI_BASE_URL/api/v1/documents/{documentId}" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"name": "Q1 Revenue Report (Updated)",
"clearExistingDraft": true
}'
Set clearExistingDraft: true if the document has an existing draft, otherwise the API returns 409 Conflict.
curl -L -X DELETE "$OMNI_BASE_URL/api/v1/documents/{documentId}" \
-H "Authorization: Bearer $OMNI_API_KEY"
Soft-deletes the document (moves to Trash).
curl -L -X PUT "$OMNI_BASE_URL/api/v1/documents/{documentId}/move" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"folderPath": "/Marketing/Reports",
"scope": "organization"
}'
Use "folderPath": null to move to root. scope is optional — auto-computed from the destination folder.
curl -L -X POST "$OMNI_BASE_URL/api/v1/documents/{documentId}/duplicate" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"name": "Copy of Q1 Revenue Report",
"folderPath": "/Marketing/Reports"
}'
Only published documents can be duplicated. Draft documents return 404.
Update the tiles, queries, filters, and visualizations on an existing dashboard using PUT /api/v1/documents/{documentId}. This is a full replacement — you must pass the complete desired state of the document, not just the fields you want to change.
Warning: This endpoint only works with dashboard documents. Workbook-only documents are not supported and return 400.
Note: This endpoint is documented at docs.omni.co but may not appear in the OpenAPI spec at
/openapi.jsonyet.
Step 1 — Read the existing document to get its current state:
curl -L "$OMNI_BASE_URL/api/v1/documents/{documentId}" \
-H "Authorization: Bearer $OMNI_API_KEY"
This returns the full document including queryPresentations, filterConfig, filterOrder, modelId, name, and other fields. Use this as your starting point.
Step 2 — Modify the response as needed:
queryPresentations arrayqueryPresentations arrayquery, config, fields, etc.filterConfig and filterOrderStep 3 — PUT the updated document:
curl -L -X PUT "$OMNI_BASE_URL/api/v1/documents/{documentId}" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"modelId": "your-model-id",
"name": "Q1 Revenue Report",
"facetFilters": false,
"refreshInterval": null,
"filterConfig": {},
"filterOrder": [],
"clearExistingDraft": true,
"queryPresentations": [ ... ]
}'
The queryPresentations array uses the same structure as document creation — see above.
| Parameter | Type | Description |
|---|---|---|
modelId | string | Model ID for query transformation |
name | string | Document name (1–254 characters) |
facetFilters | boolean | Enable facet filters on the dashboard |
refreshInterval | integer or null | Auto-refresh interval in seconds (min 60), or null to disable |
filterConfig | object | Dashboard filter configuration — pass {} for no filters |
filterOrder | array | Ordered filter IDs — pass [] for no filters |
queryPresentations | array | At least one query presentation required (same structure as document creation) |
| Parameter | Type | Description |
|---|---|---|
clearExistingDraft | boolean | Discard existing draft before updating. Required when the published document has a draft — otherwise returns 409 Conflict. |
documentMetadata | object | Presentation settings including filter collapsibility |
queryPresentation you include becomes a tile. Any tile you omit from the array is removed. Always start from the existing document's queryPresentations and modify from there.clearExistingDraft: true is set.model_extension_id).Push custom dimensions and measures to a specific dashboard by writing to its workbook model. This is a two-step flow:
Step 1 — get the document to find its workbook_id:
curl -L "$OMNI_BASE_URL/api/v1/documents/{documentId}" \
-H "Authorization: Bearer $OMNI_API_KEY"
# → response includes "workbook_id"
Step 2 — POST YAML to the workbook model:
curl -L -X POST "$OMNI_BASE_URL/api/unstable/models/{workbookId}/yaml" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"fileName": "order_items.view",
"yaml": "views:\n order_items:\n dimensions:\n is_high_value:\n sql: \"${sale_price} > 100\"\n label: High Value Order\n measures:\n high_value_count:\n sql: \"${order_items.id}\"\n aggregate_type: count_distinct\n label: High Value Orders"
}'
fileName must be "model", "relationships", or end with .view or .topic. The yaml value is a YAML string (not a JSON object). Writing to a workbook model skips git sync entirely — authorization is still checked against the underlying shared model's permissions.
curl -L "$OMNI_BASE_URL/api/v1/dashboards/{dashboardId}/filters" \
-H "Authorization: Bearer $OMNI_API_KEY"
Filters can be updated via two approaches:
PUT /api/v1/documents/{documentId} (recommended) — update filters as part of a full document update. Include filterConfig and filterOrder alongside queryPresentations and other required fields. See the Update Existing Dashboard section.PATCH /api/v1/dashboards/{id}/filters — partial filter update. Has been reported to return 405 or 500 in some configurations.For new dashboards, the most reliable way is to include filterConfig and filterOrder in the initial POST /api/v1/documents call. See references/filterConfig.md for complete examples of each filter type.
curl -L -X POST "$OMNI_BASE_URL/api/v1/documents" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"modelId": "your-model-id",
"name": "Filtered Dashboard",
"filterConfig": {
"date_filter": {
"type": "date",
"label": "Date Range",
"fieldName": "order_items.created_at",
"kind": "TIME_FOR_INTERVAL_DURATION",
"ui_type": "PAST",
"left_side": "6 months ago",
"right_side": "6 months"
},
"state_filter": {
"type": "string",
"label": "State",
"kind": "EQUALS",
"fieldName": "users.state",
"values": []
}
},
"filterOrder": ["date_filter", "state_filter"],
"queryPresentations": [...]
}'
The keys in filterConfig (e.g., "date_filter") are arbitrary IDs — they must match the entries in filterOrder. To learn the exact filter structure, read filters from an existing dashboard with GET /api/v1/dashboards/{dashboardId}/filters.
Date Range (relative) — type: "date", kind: "TIME_FOR_INTERVAL_DURATION", requires fieldName
Date Range (absolute) — type: "date", kind: "WITHIN_RANGE", requires fieldName
String Dropdown — type: "string", kind: "EQUALS", requires fieldName, values: []
Boolean Toggle — type: "boolean", requires fieldName
Hidden Filter — any filter with "hidden": true (applied but not visible)
Critical: Every filter MUST include
fieldNamewith the fully qualified field name (e.g.,"order_items.created_at"). Without it, the filter won't bind to any column. For date filters, do NOT include a timeframe bracket infieldName.
Controls change what fields or granularity tiles display. They go in a controls array (NOT in filterConfig), but their IDs are included in filterOrder.
Time Frame Switcher — type: "FIELD_SELECTION", kind: "TIMEFRAME" with options array
Field Switcher — type: "FIELD_SELECTION", kind: "FIELD" with options array
See references/filterConfig.md for complete filter and control examples.
After creating or finding content, always provide the user a direct link:
Dashboard: {OMNI_BASE_URL}/dashboards/{identifier}
Workbook: {OMNI_BASE_URL}/w/{identifier}
The identifier comes from the document's identifier field in API responses (not id, which is null for workbooks).
omni-model-explorer to find topic + fieldsPOST /api/v1/query/run (using omni-query) before building the dashboard. Include the same filters you plan to use in filterConfig as query-level filters to confirm they parse correctly. This catches field name typos, missing join paths, bad filter expressions, and permission errors before they become broken tiles.POST /api/v1/documents with queryPresentations + filterConfig + filterOrder all in one call{OMNI_BASE_URL}/dashboards/{identifier} to the useromni-content-explorer or GET /api/v1/documents to locate itGET /api/v1/documents/{documentId} to get the full document including queryPresentations, filterConfig, etc.queryPresentations array; update filterConfig/filterOrder as neededPUT /api/v1/documents/{documentId} with the complete modified document and clearExistingDraft: true{OMNI_BASE_URL}/dashboards/{identifier} to the useromni-model-builder for shared fields, or update-model for dashboard-specific fields# Start async download
curl -L -X POST "$OMNI_BASE_URL/api/v1/dashboards/{dashboardId}/download" \
-H "Authorization: Bearer $OMNI_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "format": "pdf" }'
# Poll job
curl -L "$OMNI_BASE_URL/api/v1/dashboards/{dashboardId}/download/{jobId}/status" \
-H "Authorization: Bearer $OMNI_API_KEY"