Skip to main content

Holistics Docs

Holistics is an AI analytics platform built on a uniquely expressive semantic layer.

Most BI tools now ground their AI in a semantic layer. The problem is depth: most semantic layers can only express first-order queries (slice, filter, group), so AI hits a ceiling on real analytical questions like period comparisons, cohort retention, and ratios across grains. It inherits the limits of the layer underneath.

Holistics is built differently:

  • A programmable semantic layer. Most BI tools define their semantic layer in YAML configs, which are schemaless, error-prone, and require Jinja workarounds for any reuse. Holistics uses AML, a typed modeling language where models, dimensions, measures, and relationships are first-class language constructs. Modules, extends, partials, conditionals, IDE tooling. A real language built for analytics modeling, not generic key-value configs.
  • A composable query language on top. Most semantic layers treat metrics as SQL strings that can't combine or reuse. AQL treats metrics as first-class composable objects, so period comparisons, cohort retention, and ratios across grains stay inside the metric layer instead of leaking into derived tables and spreadsheets.
  • Analytics-as-code infrastructure. Most BI tools are UI-first with Git bolted on as an afterthought (if at all). In Holistics, every definition is version-controlled in Git from day one, reviewable through pull requests, and promotable through environments. Code is the source of truth, not a side effect.

This is why trusted AI analytics, governed self-service, and embedded analytics actually work in practice. AI doesn't guess from raw tables; it reasons from the same governed business definitions your team uses. Self-service goes beyond slice-and-dice because cohort retention, period comparisons, and ratios across grains stay inside the governed semantic layer instead of leaking into spreadsheets and one-off SQL.

How Holistics fits in your Data Stack

Where to start

Understand Holistics:

Try it hands-on:

If you have a specific evaluation in mind: Holistics AI (AI on the governed semantic layer), Open Semantic Layer (query metrics programmatically), Embedded Analytics (dashboards and AI in your product), or Coming from Looker.

Stay in the loop

Older versions of Holistics

This site covers Holistics 4.0. If you're on an earlier version:


Open Markdown
Let us know what you think about this document :)