DAC - Dashboard-as-Code
A tool that lets you write a company dashboard the same way you write code — in text files you can save, review, and hand to an AI assistant to edit — instead of clicking around a chart-builder UI.
🔗 Visit DAC - Dashboard-as-Code
Description
Building a dashboard usually means dragging chart widgets around in a tool like Tableau or a BI platform, which works fine until you need to tweak it later, share exactly how it was built, or let an automated system update it. DAC treats a dashboard like a recipe written down on paper: you describe it in plain text files, and the tool reads that recipe and serves the actual charts. Because it's just text, it can live in the same place as your other code, get reviewed before it changes, and — increasingly relevant — be written or edited by an AI agent instead of a human clicking through menus.
Technically, DAC defines dashboards in YAML (for structure — tabs, loops, conditionals) and TSX (for dynamic chart components), with a built-in semantic layer that turns named business metrics into SQL automatically generated against your warehouse — Postgres, MySQL, Snowflake, BigQuery, Redshift, or Databricks. It's AGPL-3.0 open source, built by Bruin (a data pipeline company) and integrates with the Bruin CLI, with the explicit design goal of being something an AI agent can generate and modify reliably, not just a human-authored config format.
💬 Our review
The short version: if your team already treats infrastructure and pipelines as code and dashboards are the one thing still living in a point-and-click UI, DAC closes that gap for free — the catch is you're trading visual convenience for a YAML/TSX learning curve.
Against Grafana or Metabase, DAC's edge isn't the charts themselves (those tools render dashboards fine) — it's that a dashboard becomes a diffable, version-controlled artifact and a target an AI agent can safely generate without hallucinating a UI click-path. Against Evidence.dev, which pioneered dashboard-as-code with Markdown+SQL, DAC's semantic layer is the differentiator: define a metric once and every chart references it, rather than re-writing SQL per dashboard. The real risk is adoption stage — AGPL-3.0 means any hosted redistribution of a modified version has to stay open, which matters if you're a vendor, and the semantic layer commits you to Bruin's opinions about how metrics should be modeled. Fine for a data team building internal tools; less fine if you want the flexibility to switch BI vendors later.
📊 Global score
🤖 AI-enriched data
AGPL-3.0, entièrement gratuit et self-hosted. Aucun coût logiciel — seuls les coûts d'infrastructure (base de données, hébergement) s'appliquent.
Pros
Dashboards en texte versionnable (YAML/TSX)
Couche sémantique intégrée pour les métriques
Support des principaux entrepôts de données
Pensé pour être généré/édité par des agents IA
Cons
Pas d'éditeur visuel drag-and-drop
Licence AGPL-3.0 contraignante en redistribution hébergée
Dépendant des conventions Bruin
Écosystème plus jeune que Grafana/Metabase