A value read down a scale, like a depth profile. Sample data: recall of Atlas 3 and Kite 2 Pro from 4K to 128K context, on a 60% to 100% scale.
Installation
- 1
Set up the registry once
This chart is free in Hearth and MIT licensed: no token. Add @chartory to components.json once (the URL alone, no headers), or skip this when another chart is already installed.
Installation guide - 2
Add the chart
One command adds the Vertical profile component to your project.
npx shadcn add @chartory/vertical-profile - 3
Use it
Pass your rows. The props are the Figma properties.
usage.tsximport { VerticalProfile } from "@/components/chartory/vertical-profile"<VerticalProfile data={recallByContext} size="m" legend/>
Props
<VerticalProfile>
Every prop but data is a Figma property with the same name and values.
VerticalProfileData
The data prop: one object for the whole chart.
VerticalProfileSeries
One object per entry in series.
The data as a JSON Schema: schemas/vertical-profile.json
Vertical profile chart: a React and Figma chart in four sizes
What does a vertical profile chart show?
A vertical profile chart plots a value against a scale that runs up the side, such as context length, altitude or depth, so you see where performance holds and where it falls off. Use it for model recall by context length or any measure read at levels.
Chartory ships it as a Figma component and a React component with the same name. The code installs from the shadcn registry with one command and lands in your project as a file you own.
What is the Vertical profile component?
It is one <VerticalProfile> component. Its props are the Figma properties with the same names and values, and the props table above lists them. The data is a value per level, one or two series.
Sizes XS, S, M and L re-lay the chart out from the data instead of scaling it, so labels stay readable at every size.
When should I use it?
- Is the scale levels, not time? Yes means the vertical profile fits.
- Are there one or two series? Yes means the vertical profile fits.
- Is it a profile, not a spread? Yes means the vertical profile fits.
When does another type fit better?
- Values over time. Use Multi-line chart instead.
- Many models across measures. Use Parallel coordinates instead.
- A spread of values. Use Ridgeline instead.
Vertical profile questions
Installing, data, Figma and plans.
Yes. It installs through the shadcn registry with the shadcn CLI, and the file lands in your project.
It is free in Hearth and MIT licensed: add the @chartory registry URL once, no token, then run npx shadcn add @chartory/vertical-profile. The Installation guide has the steps.
Yes. Connect the Chartory MCP server and ask for the vertical profile chart; the agent runs the same command and you review the diff.
A value per level, one or two series. The sample data comes with the component, and the props table lists every field.
XS, S, M and L, in Figma and in code. Each size is laid out again from the data, so text and lines keep their size.
Yes. chart/016-vertical-profile has the same properties as the component.
Yes. Every paint is a token, so the chart follows light and dark mode and the collection you pick.
None for Hearth: it is free in the public registry, no token. The Signal collection (@chartory/signal) comes with Code and Complete.