AG Studio: Embedded Analytics for Web Applications

AG Studio is an embedded analytics component that sits inside your own application, built on AG Grid and AG Charts. Once a developer has embedded it, your end users can build and customise dashboards and reports directly within your app, without needing a developer for every change. See the Overview.

Neither. AG Studio is a ready-made front-end component that you embed into your existing application and connect to your own data. Your application keeps ownership of the surrounding shell: routing, authentication, data fetching and persistence.

Three layers: a drag-and-drop dashboard canvas with panels for editing widgets, data and filters; edit and view modes, which your application switches between to control who can build and who can only consume; and a built-in data engine that handles aggregation, filtering, cross-filtering and calculated fields automatically. See Modes & Layout.

Building dashboards in-house is slow and expensive to maintain, while sending users to an external BI tool, or embedding a rigid third-party one, breaks the in-app experience and limits customisation. AG Studio embeds directly inside your application, gives your end users self-service dashboard building, and processes your data in the browser or through your own backend, never through AG Grid.

No. AG Studio isn't a separate tool you send users to. A developer integrates it into your application and makes your data available to it. Once set up, your end users can self-serve from within your app.

AG Studio is built on a framework-agnostic core, with official wrappers for React, Angular and Vue 3, and it also works with plain JavaScript or TypeScript. If you're already using AG Grid in one of these frameworks, Studio fits your existing setup. See Installation.

Yes. It handles aggregation, filtering and cross-filtering between widgets, relationships between multiple tables, and calculated fields out of the box, without needing a separate query layer. For large datasets, you can replace it with a Custom Engine that sends each query to your backend instead of computing in the browser.

No. AG Studio works with any data source, such as PostgreSQL, BigQuery, Snowflake, ClickHouse or a REST API, but it doesn't ship native database connectors. Your application provides the connection, either by fetching the data itself or through a Custom Engine.

The built-in data engine handles up to around a million rows in the browser, though the practical limit depends on how complex your calculations are and on the end user's machine. For larger or more complex datasets, a Custom Engine moves the computation to your backend.

AG Studio is a front-end component, so it doesn't connect to a database directly; it gets data through your code. Either your application fetches rows and passes them to Studio, which then processes them in the browser, or you provide a custom engine that translates Studio's queries into SQL or API calls against your backend. See Loading Data.

Yes. You implement a Custom Engine that replaces the built-in engine, so your backend executes each query and returns the results to Studio.

Yes. AG Studio supports multiple tables linked by defined relationships, similar to a semantic layer in traditional BI tools, so fields from different tables can appear together in the same widget. See Data Modelling.

Yes. AG Studio uses the same theming API as AG Grid and is highly customisable. There's also a Theme Builder and a Figma Design System. See Theming.

Yes. You choose which panels appear in edit and view mode and on which side, and you can customise what the Edit Panel shows, including which widgets users can add. See Modes & Layout.

Yes. AG Studio provides translations for a range of commonly requested languages, lets you override any string, and can hook into your application's own localisation system. See Localisation.

Yes. A Custom Widget can render any component, including your own components, other charting libraries, maps or diagrams, alongside Studio's built-in widgets.

No. The built-in table widget is a purpose-built AG Grid with limited interactivity; for example, columns can't be reordered by dragging, since drag-and-drop is reserved for the canvas. If you need full AG Grid functionality, you can add a complete AG Grid as a Custom Widget. AG Grid Enterprise features in a custom widget need their own AG Grid Enterprise licence.

Yes. A developer can build a report in edit mode, save its state and ship it with your application, so users open a ready-made report. Load it in view mode for a fixed report, or in edit mode as a starting point users can change. See State and Modes & Layout.

Yes. Dashboards export to PDF through the browser's print dialogue, using a print-optimised layout. Chart and grid widgets export to CSV (up to 1,000 rows by default, which you can change), and chart widgets can also be downloaded as an image. See Exporting.

Yes. Users can open a detail report filtered to the item they right-clicked, such as a row in a table. Your application chooses which filters carry over and where the report opens: in place, in a popup or anywhere else. See Drill Through.

Not currently. There's no built-in commenting, annotation or real-time collaboration. Dashboard state is a single JSON object, though, so your application can let users save a dashboard and share it with a colleague. See State.

Yes. The Agent Framework lets users explore data and build or change dashboards in natural language, through a built-in chat panel connected to your own LLM.

The conversation, the agent's instructions and your field and widget names, plus the results of any queries the AI runs. It goes only to the provider you connect, never to AG Grid. See Agent Framework Overview.

They are off by default, and the chat panel only appears once you enable them. See Agent Quick Start.

Any. AG Studio doesn't include a model; you connect your own provider, such as OpenAI or Anthropic. See Harness Overview.

AG Grid adds no per-query charges. You pay your LLM provider directly, at its own rates.

Yes, it is fully self-hosted. AG Studio runs entirely within your own application and infrastructure and makes no calls to AG Grid's servers, so AG Grid has no visibility into your deployment, data or usage. The only outbound request it can make is loading Google Fonts for a theme, and only if you opt in. See Security.

AG Studio captures the whole report, including its pages, widgets, layout and filters, as a single JSON object. Your application saves it wherever it chooses, such as your own database, and loads it back to restore the report. Sharing a report is then a matter of giving a colleague access to its saved state. See State.

Yes, through your own permission system. Your application supplies the data and field definitions to Studio, so it can tailor them per user or role. To restrict a field, leave it out of the data you pass in, or out of what your custom engine returns. See Loading Data.

Yes, as an application-level decision. Your application can expose a public page with Studio in view mode: viewers can still filter the dashboard, but nothing is saved unless your application chooses to persist it. See Modes & Layout.