Build production dashboards with AI — safely
You describe what you want in plain language; your AI assistant builds it behind three layers of guardrails that keep it from inventing APIs, UI, or insecure backends. You don't have to become an engineer to ship a real dashboard.
Your first project
Dashboards come in two frontend flavours: React and Angular — pick whichever your team uses. The Quickstart has step-by-step commands for both.
React (with a full-stack backend). One command scaffolds the React frontend and the Flask backend, wired together under one shared dashboard id:
npx @rystad/create-rystad-dashboard my-dash --dashboard-id 300
This is the umbrella project: not just a frontend, but the whole stack you
run together behind the real gateway with docker compose up. On the first run
the CLI installs both sides and starts the stack for you once the prerequisites
are in place.
Angular. Prefer Angular? Scaffold a modern Angular dashboard (standalone components, signals, TanStack Angular Query) instead:
npx @rystad/create-rystad-angular-app my-dash --dashboard-id 300
Same guardrails and MCP guidance, tailored to Angular. It's frontend-only today
(pair it with @rystad/create-rystad-api if you need a backend).
New here? Begin with First-time setup — it authenticates you
to the private @rystad feed (skip it and the command fails with an opaque
401, or a 404 if you drop the @rystad/ scope).
Find your way around
- Getting started — Why the system exists, First-time setup, and the Quickstart.
- Learn — the Concepts behind the guardrails and Safe prompting.
- Reference — generated docs for each MCP server (endpoints, recipes, guardrails).
- Help & examples — the FAQ and the Gallery.