portaljs: an AI-native framework for building data portals with agentic architecture advisory and Next.js scaffolding
portaljs: an AI-native framework for building data portals with agentic architecture advisory and Next.js scaffolding
What it solves
PortalJS provides an AI-native framework for building data portals. It eliminates the complexity of deciding on data infrastructure (storage, compute, and cataloging) and the manual effort of wiring a frontend to that data, allowing teams to move from a brief to a a working portal quickly without being locked into a proprietary platform.
How it works
PortalJS uses an agentic skills framework—specifically designed for Claude Code—to guide users through the architecture and build process. It consists of three main components:
- Agentic Skills: A set of slash commands (e.g.,
/portaljs-architect,/portaljs-new-portal) that advise on the best infrastructure stack and scaffold plain Next.js code. - Three Surfaces: A standardized layout consisting of a Home page, a Catalog for discovery, and a Showcase for exploring individual datasets.
- DataProvider Contract: A pluggable backend system that allows the portal to switch between static files, CKAN, OpenMetadata, or an "open lakehouse" (Git + R2 + Parquet + DuckDB) without changing the frontend code.
Who it’s for
Data teams and organizations that need to build professional, scalable data portals to share and explore datasets, particularly those favoring open-source tools and serverless architectures (like Cloudflare R2/Workers).
Highlights
- AI-Driven Architecture: Includes an advisor skill that recommends a storage and compute stack based on the user's specific data needs.
- No Lock-in: Generates plain, editable Next.js code that the user owns entirely.
- Open Lakehouse Default: Promotes a modern, serverless stack using Parquet and DuckDB over traditional heavy data warehouses.
- Pluggable Backends: Supports a wide range of sources including CKAN, GitHub, and Frictionless Data Packages via a single data provider contract.
- Extensive Tooling: Provides specialized skills for adding datasets, creating charts/maps, defining schemas, and migrating data from ArcGIS Hub or Socrata.
Sources
- undefineddatopian/portaljs