Embedded analytics that feels like your product

Embedded analytics works by connecting an analytics layer to your data, defining what each customer may see, and rendering the result inside your product. With Databrain, you connect your warehouse with read-only access, model a datamart with your tenancy rules, and build dashboards in the builder. You then add a web component such as dbn-dashboard to your app, your backend requests a guest token for each user, and every query runs live against your warehouse, scoped to that customer.

Databrain scopes every embedded session to one tenant through a guest token your backend requests from the Databrain API with that tenant's ID. For metrics built in the visual builder, the tenant filter is injected into the generated SQL at query-build time; custom SQL metrics reference an explicit tenant variable. A request from the browser cannot widen the scope the token was issued for. Databrain supports database-per-tenant, schema-per-tenant and shared-table tenancy models.

Databrain embeds with web components, dbn-dashboard and dbn-metric, which render inside a Shadow DOM in your page instead of an iframe. Dashboards sit in your layout and are styled through Databrain themes and component options, isolated from your page's CSS. Databrain publishes guides for React, Next.js, Vue, Angular, Svelte, Solid and vanilla JavaScript; Angular, Vue, Svelte and vanilla hosts add react and react-dom as dependencies. An iframe option exists but is documented as not recommended.

Databrain dashboards load as fast as your warehouse answers the underlying queries, because each chart query is pushed down and runs live against your database. To keep repeat loads fast, Databrain can cache query results per workspace with an expiry you set, in its managed Redis or in your own Redis or ElastiCache instance. For large tenants, warehouse design such as sort keys, partitioning and a read replica for Postgres remains the biggest lever.

To embed Databrain you need a connected data source, at least one dashboard, a decided tenancy model, and every domain your app runs on, including localhost and staging, added to the Data App's allow list. Your backend also needs an endpoint that requests a guest token from the Databrain API for each signed-in user; API and app tokens must never reach the browser. The quickstart walks through a sample embed before you wire up production.

The most common example of embedded analytics is a customer-facing reporting module inside a B2B SaaS product, where each customer sees dashboards on its own activity without leaving the app. Databrain customers show the range: SpotDraft added customer-facing reporting to its contract platform, Freightify built a logistics analytics module, Spendflo replaced in-house analytics on its procurement platform, EpochOS gives mortgage brokers natural-language insights, and BerryBox delivers risk analytics to HR teams and insurance brokers.