Charticulator Is Archived: How to Migrate Your Power BI Charts (2026)

The Charticulator visual inside Power BI is a compiled custom visual. Retirement did not delete it from reports; it stopped maintenance. That means three things in practice.

Existing reports keep rendering until a Power BI change breaks them. There is no fix when that happens. Users began reporting exactly this in 2024: a Fabric Community thread from September 2024 describes a Charticulator visual that "is gone - just blank area where the visual was placed, all fields gone."

New designs are harder to make. The charticulator.com designer no longer receives updates, and the "export as Power BI visual" flow depended on the app being maintained.

Tenant policy can remove it overnight. Charticulator was certified, so it survives a certified-only policy, but an admin disabling the specific visual in the organisational store makes it disappear from every report at once.

What does not break: your chart definitions. Charticulator saved designs as .chart files and templates (.tmplt), and the data binding lives in Power BI. Both survive migration.

Open each Charticulator visual in Desktop while it still renders and export the design from its editor (the Export menu offers the chart file and a template). Save the .chart file, a screenshot, and a note of which fields fed each data role. If the visual has already gone blank, the design is still inside the report: save the report as a Power BI project (.pbip), find the visual's folder under definition/pages and read its visual.json. The Charticulator template is stored in the visual's properties as a long string you can copy into a file.

Also write down behaviour that is easy to forget: sort order, which axis the legend drove, conditional colour rules, and tooltip fields. A rebuild that reproduces the shape but loses the sort is a regression your readers will notice.

Charticulator was open source, and a long-time contributor, Ilfat Galiev (zBritva on GitHub), took the code forward. His fork ships as two Microsoft Marketplace visuals, "Charticulator Visual Community (Editor)" and "Charticulator Visual Community (View)", with the designer hosted at ilfat-galiev.im/charticulator and the source at github.com/zBritva/charticulator-visual-community. The Editor visual lets you design inside the report; the View visual renders a finished template.

This is the right path when you have many Charticulator charts and little time: templates created in the original tool load in the fork, so migration is mostly re-importing and re-binding fields. The caveats are the ones that apply to any single-maintainer project. Check the Marketplace listing's last-updated date and certification status before you standardise on it, and keep your exported templates, which you now know how to recover.

Deneb is free, MIT-licensed and Microsoft-certified, and it renders anywhere the report does, including Publish to web and Report Server. Most Charticulator designs are expressible in Vega-Lite because both tools think in marks and encodings: a glyph in Charticulator is a mark in Vega-Lite, a plot segment is an axis or a facet, and a constraint usually maps to an encoding channel or a transform.

The honest cost is the grammar. Vega-Lite is JSON you write and debug; the Deneb alternatives guide linked at the end weighs that against the other options. For a one-off chart it is a few hours. For a design system of twenty charts, consider whether the community fork or a generated visual gets you there faster.

A surprising share of Charticulator charts were built because, at the time, a native visual could not do one thing: a reference label, a rounded bar, a small-multiple layout, data labels inside bars, a dumbbell. Several of those gaps have since closed. Before rebuilding, check whether the current native bar, line, scatter, matrix or the new card visual (generally available since November 2025) can do the job with conditional formatting and a reference line. Native wins on performance, accessibility and zero governance friction, so take it whenever it fits.

When the chart is custom in a way no native visual reaches, and you do not want to adopt Vega-Lite or depend on a community fork, the remaining option is a proper custom visual built with the Power BI visuals SDK: TypeScript, D3, a formatting pane, selection and tooltips, packaged as a .pbiviz. Historically that meant hiring a developer. VizForge generates the four source files from a plain-English description (plus your screenshot of the old chart), type-checks them with the real TypeScript compiler, packages the visual and gives you the .pbiviz and the code.

Trade-offs, stated plainly: the result is an uncertified custom visual, so your tenant must allow SDK visuals; it costs credits (1 to 5 per visual, one-time); and you iterate through prompts and a refinement loop rather than dragging constraints. In return you own the code, nobody can retire it, and the formatting pane is designed for your chart rather than generic.

A common Charticulator design was a ranked lollipop: one row per product, a thin line from zero to the value, a circle at the end, and the rank number as a label on the left. Here it is in Vega-Lite for Deneb. Bind a table with the fields Product and Sales to the visual, then paste the spec. The rank is computed inside the spec with a window transform, so you do not need a DAX ranking measure.

The equivalent VizForge prompt reads: "Ranked lollipop chart. One row per category, sorted descending by the measure. Draw a 2 px line from zero to the value and a 7 px circle at the value. Show the rank number left of the category label. Cross-filter on click, tooltip with category and formatted value, formatting pane for colours and font size." Either path reproduces the Charticulator chart; Deneb is free and certified, the generated visual is standalone and owned.

Many charts, little time, happy with a community-maintained visual: Path A. Need Publish to web or Report Server, and someone on the team will learn Vega-Lite: Path B. The chart turns out to be reproducible with conditional formatting: Path C, always. A design that is genuinely custom and that you want to own outright without learning a grammar: Path D. Whatever you pick, keep the exported .chart files; they are the specification your rebuild is tested against.