AGX is a desktop and web-based application for exploring and querying data through a modern, user-friendly interface. Built with Tauri , SvelteKit, and Plot, AGX combines the performance of a native desktop application with the flexibility of a web interface. It enables users to interact with data stored in ClickHouse and Apache Iceberg, offering features like an interactive SQL query editor, schema browsing, and tabular result displays. The public repository is available at https://github.com/agnosticeng/agx .
Purpose
AGX provides a powerful, AI-enhanced analytics platform for querying and visualizing data. It supports both local and remote data sources, making it ideal for analysts and data engineers working with large datasets. Key features include:
- Native and Web Modes: Runs as a native desktop app on macOS and Linux using a local ClickHouse instance, or as a web app at https://agx.app for remote server access.
- Interactive Querying: Offers a SQL editor with syntax highlighting and LLM integration for assisted query writing.
- Schema Exploration: Includes a schema browser to navigate data structures.
- Cross-Platform Support: Compatible with macOS, Linux, and web browsers, with drag-and-drop file operations and tabular result views.
Integration with AGP
AGX integrates seamlessly with the Agnostic Asynchronous Query Processor (AGP) to execute queries against ClickHouse clusters. Users can connect AGX to an AGP instance by passing the AGP server URL as a query parameter in the web version’s URL, e.g., https://agx.app?proxy=https://your-agp-server. This allows AGX to submit asynchronous SQL queries via AGP’s API, track execution status (PENDING, RUNNING, SUCCEEDED, etc.), and retrieve results from AGP’s object store, leveraging AGP’s deduplication and tier-based resource allocation for efficient query processing.
Hybrid Application with Tauri
AGX is a hybrid application built using Tauri , a framework for creating lightweight, secure, and cross-platform applications with web frontends. Tauri enables AGX to run as a native desktop app on macOS and Linux, utilizing the system’s webview for rendering, with a Rust backend for performance and security. Alternatively, AGX operates as a web application, providing flexibility for users to access it via browsers without installation. This hybrid approach ensures minimal resource usage while supporting modern web technologies like SvelteKit for the frontend.
AGX’s integration with ClickHouse, AGP, and the datasets from https://github.com/agnosticeng/datasets makes it a versatile tool for data exploration and analytics, suitable for both local and cloud-based workflows. For more details, refer to the AGX repository or the AGP documentation.