01Metric layer
Write SQL or Python, get a governed metric
Define business logic once and reuse it across dashboards, APIs, and AI answers with tenant policies applied before every query.
For engineering teams
From one SDK to a complete analytics experience.
Embed tenant-scoped dashboards, metrics, and AI with components built for your stack.
- Read-only warehouse
- Tenant-scoped tokens
- SOC 2 Type 2
- ISO 27001
Build customer-facing analytics for the modern data stack
Databrain embeds white-labeled dashboards, self-service reporting and governed AI chat into your product, all built on a warehouse-native, API-first foundation.
Connect your warehouse
Point Databrain at Snowflake, BigQuery, Redshift, or any of the 18 sources you already run. Read-only, metadata-only, nothing copied.
Embed analytics anywhere
Add dashboards, standalone metrics and governed AI chat directly into your product with typed web components.
Scheduled reports, not just dashboards
Send reports on the cadence your customer picks, emailed from your own subdomain through your own SMTP server.
- API
Manage with APIs and guest tokens
Provision workspaces, mint tenant-scoped guest tokens, and manage embeds through documented REST APIs.
Build, govern, and deploy from one workflow
Scroll through the architecture your team ships, from the first metric definition to an enterprise-ready deployment.
02Grounded AI
Governed AI chat, not a black box
Every answer resolves against your semantic layer, cites the metric context it used, and declines unsupported questions.
03Model routing
Bring your own LLM
Route governed analytics requests through OpenAI, Claude, Azure OpenAI, Llama, Mixtral, or Gemini without rebuilding the experience.
04Managed cloud
Deploy in Databrain Cloud
Use a fully managed service with monitoring, scaling, and continuous product updates handled by Databrain.
05Private infrastructure
Deploy self-hosted
Run the same product inside your VPC when data residency, private networking, or infrastructure control is required.
06Security perimeter
Security and trust, built into every layer
Tenant isolation, read-only warehouse access, auditable queries, and verified controls protect the complete analytics path.
A developer platform your team can extend
Move from your first embed to governed, AI-assisted analytics without switching tools.
Typed Embed SDK
Typed Embed SDK and smart chart runtime
Embed dashboards or standalone metrics through a stable web-component contract. Cross-filtering, drill-down, paged tables, date controls, and dark mode, without forking the analytics UI.
Semantic layer
Author metrics once, reuse everywhere
Model metrics, measures, joins and policies. Give every dashboard, API, and AI assistant the same business definitions to work from.
MCP
MCP for AI coding assistants
Connect Claude, Cursor, or Windsurf through MCP. The same grounded tool runtime lets an assistant suggest SQL-backed metrics and chart configurations for your review, Databrain never writes back to your systems unsupervised.
API
APIs and workflow hooks
Provision and manage analytics through documented REST APIs, and send insights into the operational tools your team already uses.
Framework agnostic
Use the framework your product already runs on
Use the same typed embed contract across your frontend stack. Change frameworks without rebuilding the analytics layer.
React
Next.js
Vue.js
Angular
Svelte
JavaScript
Warehouse-native
Connects to the warehouse you already run
One read-only connection. Live queries keep your warehouse as the source of truth.
- Read-only connection
- Live queries
- No data copied
Questions engineers actually ask
What is an embedded analytics SDK?
An embedded analytics SDK is a developer package that renders dashboards and charts inside your own application, with authentication, tenant scoping and theming handled by the SDK and the vendor’s backend rather than code you write. Databrain’s SDK is the @databrainhq/plugin npm package: it registers the dbn-dashboard and dbn-metric web components, which render in Shadow DOM so your CSS and Databrain’s do not collide, and authenticate with a guest token your server creates.
How do I embed Databrain dashboards in React or Angular?
Install @databrainhq/plugin, import @databrainhq/plugin/web once, and render the dbn-dashboard component with a dashboard ID and a guest token. Your backend creates the token by calling POST /api/v2/guest-token/create with your API token and the customer’s clientId, so the API token never reaches the browser. Databrain publishes guides for React, Next.js, Vue, Angular, Svelte, Solid and vanilla JavaScript. Non-React hosts such as Angular and Vue should add react and react-dom as dependencies.
How does authentication work for embedded Databrain dashboards?
Databrain uses guest tokens: opaque, server-issued tokens, not JWTs. Your backend holds the API token and calls the guest token endpoint with the customer’s clientId, plus optional row-level rules, filters and an expiry; the browser only ever receives the guest token. Databrain checks embedded queries against that token’s scope, so a clientId changed in the browser cannot widen access. Guest tokens are reusable until they expire, so cache them per user on your backend.
Is there a Python SDK for Databrain?
There is no separate Python SDK. Databrain’s server-side surface is a REST API you can call from any backend language, and the docs include Python examples, such as creating a guest token from a FastAPI endpoint. The frontend embed is the JavaScript web component package. Python also works inside Databrain itself: the metric builder includes a Python editor for writing metrics in code alongside SQL.
What can I automate with the Databrain API?
Databrain’s REST API covers the setup work you would otherwise click through: creating workspaces, data sources, datamarts and Data Apps, creating and updating embeds, importing and exporting dashboards, managing the semantic layer, rotating API tokens, whitelisting domains and creating guest tokens. For AI coding assistants, Databrain also ships an MCP server with 36 tools; its production write tools default to a dry run where supported and require explicit confirmation before applying changes.
How do I theme embedded Databrain dashboards?
Databrain dashboards take a theme-name that matches a theme saved in Databrain’s UI theming settings, or theme and admin-theme-options objects passed inline on the component. Those cover fonts, dashboard and card colors, chart palettes and date formats, while per-embed attributes control options such as downloads and full-screen behavior. Labels can be translated per user with the language and translationDictionary props, and there is no Databrain badge in the embed to remove.
Start with the quickstart, not a sales call
Five minutes to a rendered dashboard. If it fits, bring your product lead to the demo afterwards.