Open source visualisation build tool for the data pipeline.

glyf lets analytics engineers define charts in SQL, next to the models those charts read from. Chart definitions live in .ggsql files, move through code review like any other code, and compile into versioned artifacts — PNG, SVG, Vega JSON, and static dashboard sites — with no BI server anywhere in the loop.

glyf is artifact-driven, not runtime-driven. It reads the metadata your transformation tool already produces — starting with dbt's manifest.json, resolving ref() and source() from it — executes the chart SQL, and writes deterministic output you can inspect, diff, archive, and publish anywhere.

SELECT week, plan, sum(activated_users) AS activated_users
FROM {{ ref('fct_product_usage') }}
GROUP BY 1, 2

VISUALISE week AS x, activated_users AS y, plan AS color
DRAW line
LABEL title => 'Activation by Plan'
INTERACT tooltip, legend_filter

dbt build      # your models, unchanged
glyf build     # charts, dashboards, static site
glyf serve     # local preview

The chart syntax builds on ggsql, which brings Grammar of Graphics clauses — VISUALISE, DRAW, SCALE, LABEL — into SQL workflows.

uv tool install glyf-core   # PyPI package is glyf-core; the command is glyf
glyf --version

Wheels ship for Linux, macOS, and Windows on Python 3.11+. Other options are in the installation guide.

Then follow the Quickstart to scaffold your first chart in a dbt project you already have. The dbt integration guide covers how glyf reads dbt's artifacts.

• Documentation — https://glyfdata.com/docs/intro
• Examples gallery — https://glyfdata.com/docs/examples/gallery
• Roadmap — https://glyfdata.com/docs/resources/roadmap

• Slack — join the glyf community. The quickest way to get help; ask in the help channel.
• Discussions — https://github.com/glyf-data/glyf/discussions
• Issues — https://github.com/glyf-data/glyf/issues

Pull requests, bug reports, and thoughtful issues are welcome. Start with CONTRIBUTING.md, and see GOVERNANCE.md for how decisions get made.

Projects in this organisation follow the Contributor Covenant Code of Conduct, version 2.1.

Licensed under Apache License 2.0.

Please report security vulnerabilities privately through GitHub security advisories rather than in a public issue.