Headless BI refers to a business intelligence software that operates without a traditional user interface. In other words, it is a BI product that delivers data and insights to other applications or systems, rather than presenting them to end-users through a visual dashboard or reporting tool. In fact this is a special case of embedded BI where data visualisation part is custom but reporting API is provided by the headless BI component. Main benefits are:
On practice headless BI product is a self-hosted server application that exposes reporting web API for making analytical queries and data exports. This application has a configuration (usually in the form of JSON or yml) that contains all necessary metadata needed to translate these analytical API queries into database queries.
Product | Price | Pros | Cons |
---|---|---|---|
Cube.js |
Core: free
Cloud: $$
Premium: $10k/year
|
Open source / Free when self-hosted ("Cube Core"). Built-in pre-aggregations. Client-side API library (JS), integration examples. A lot of built-in data connectors. | No high-level API for reports (only raw data is returned). No built-in exports to PDF, Excel, CSV or rendering to HTML/image. No XMLA connector (for SSAS). |
MetricFlow |
(free)
|
Free/Open source (self-hosted). This is more a semantic layer middleware which hides complexity of underlying DB queries from client applications. | Non-aggregated queries are not supported. MetricFlow doesn't offer any API for reports generation. This means that all HTML rendering/exports should be implemented on your side. Introduces its own MQL syntax. Lack of support by many BI tools. Many DBs are not supported either (like MySQL). |
PivotData Microservice |
$
Starts from $999
|
High-level web API for tabular reports HTML generation, including pivot tables. All exports are built-in (PDF, Excel, CSV, JSON and HTML output). Web API is suitable for end-user report builder (client-side example is included). A lot of built-in data connectors (including XMLA!). Supports custom connectors (your own web API as a data source), so it is possible to use as a sematic layer data sources like cube.js / metricflow. | No built-in pre-aggregations (short-living in memory caching is present). |
GoodData |
$$
Starts from $12,000/year
|
Combines embedded BI capabilities with API-first approach. Offers predictive analytics capabilities, strong security and governance features. | Cloud-native solution (free self-hosted Community Edition is suitable only for evaluation), enterprise-grade pricing. |
icCube |
$$$
More than $22,000/year
|
This is an OLAP server with XMLA/MDX interface with a separate "Web Dashboards" front-end / Supports Excel as a client / can be used as a drop-in SSAS replacement | To benefit from the full power of icCube you need MDX knowledge. |
PowerBI |
$$$
Premium starts from $59,940/year
|
PBI has a sematic layer inside and it can be used as a 'headless BI' via Execute Queries API / Reports Export To File API or XMLA endpoint (so even Excel can be used as a front-end!). | XMLA endpoints / self-hosted version require PowerBI Premium which is rather expensive. No export to simple <table> HTML. |