AGENT SKILLChart selection, generation, and styling.

OPEN-SOURCE VISUALIZATION FOR AI

Turn AI answers into charts. AI-Native Visualization for the LLM Era.

Ask your assistant for a chart, or bring live visualizations into your app. Render 26 visualization types as your model responds, from statistical charts to relationship graphs, tables, and summaries. Style them all with one theme.

Use with your agent Build with the SDK

Use in desktop assistants, coding agents, and editors that support Skills.

26 chart typesSyntax & JSONAgent Skill includedBuilt on G2 · G6 · T8

01 / GET STARTED

Your agent.
Now with charts.

Ask your agent to pick a chart, generate its configuration, or build a reusable theme. Or bring the rendering SDK into your application.

YOUR MODEL. YOUR WORKFLOW.

INSTALL ONCE. KEEP CREATING.

Start in your assistant.

Install the Skill, then give your assistant its first chart task.

DESKTOP & PERSONAL ASSISTANTS

ChatGPT desktop

Claude

QwenWork

WorkBuddy

QoderWork

OpenClaw

CODING AGENTS

Codex

Cursor

Claude Code

Gemini CLI

Copy the installation prompt and paste it into your assistant to get started.

Read the prompt

Install GPT-Vis's chart-visualization skill for the assistant I am using.

Source: https://github.com/antvis/GPT-Vis/tree/ai/skills/chart-visualization

Use this assistant's supported skill installation location. Download the complete chart-visualization directory from the ai branch, including SKILL.md and all references. If using the skills installer, select only chart-visualization.

Read the installed SKILL.md, verify that its reference files are present, and report the installation location and whether a new session is needed. If this assistant requires a manual upload or cannot install from a link, explain that instead of claiming success.

If your assistant cannot install from a link, download the ZIP and import it in its Skills settings.

Download Skill ZIP

Upload steps & terminal install

WorkBuddy: Skills → Add Skill → Upload Skill.

Claude Chat / Cowork: Customize → Skills → Create skill → Upload a skill. Enable it after upload.

Terminal: Run in your project folder, then choose your agent.

npx skills add https://github.com/antvis/GPT-Vis/tree/ai --skill chart-visualization

Client setup guides Skill source

Make your first report.

Try a monthly revenue report with a line chart, a table, and one shared theme.

Jan$12,000Feb$18,000Mar$15,000Apr$24,000

Read the task

Use the GPT-Vis chart-visualization skill to create a monthly revenue report from these values (USD): Jan 12000, Feb 18000, Mar 15000, Apr 24000.

Include a line chart, a table with the same four values, and a short summary of the total revenue and April's change from March. Calculate the figures from the supplied data.

Use a shared theme configuration for the chart and table. Create report.html that I can open directly in a browser, with the theme embedded and any published library dependencies loaded. Also save the same theme configuration as theme.json so I can reuse it. No build step should be needed.

02 / BUILT FOR MODEL OUTPUT

The answer streams in.
The chart takes shape.

Keep passing the accumulated output to render(). GPT-Vis updates the chart as more data arrives. Switch between Syntax and JSON below to compare the two formats.

“Show me how revenue grew in the first half of the year.”

EXAMPLE

agent output

01vis column02title Revenue ($k)03data04  - category Jan05    value 18

Taking shape…1 / 6 data points

Same data. Two readable formats.

This is a simulated response using fixed sample data. Read the streaming guide →

03 / ANALYSIS, MADE VISIBLE

Ask a question.
See the evidence.

AVA already uses GPT-Vis to turn analysis results into charts. Bring that workflow into your agent or application, with the underlying values and chart configuration kept in reach.

SAMPLE DATA / LIVE CHARTS

Which region brings in the most revenue?

THE VISUAL ANSWER

Revenue by region

GPT-Vis

Total revenue: $125,287
East’s share: 36.2%

USD

THE TAKEAWAYEast leads with $45,382, contributing 36.2% of total revenue.

WHAT’S NEXT COMING SOON

Give each finding more room to speak.

Next, we’re adding ways to mark a finding, show a statistical range, and select data for further analysis in your app.

Make the point.

Highlight a value, add a target line, or mark the period that matters.

Annotations · Coming soon

Show the range.

Display precomputed statistics, error bars, and bands alongside your data.

Statistical expression · Coming soon

Explore the detail.

Compare groups in facets and return selections to your application.

Facets & selection · Coming soon

APPEARANCE / MAKE IT YOURS

Charts that match
your style.

GPT-Vis supports theme configuration for colors, fonts, and chart styles. Try a gallery preset below, or create a theme from your brand, slides, or UI in Studio.

LIVE PREVIEW

Q4 / BUSINESS REVIEW

The same report. Your style.

GPT-Vis

Statistics

Relationship

Table

Summary

Start with a gallery style, then make it yours.
Import JSON, adjust your theme, and preview all 26 chart types in Studio.

Download the full theme JSON to use a gallery style in your app.

BUILT IN THE OPEN · AN ANTV PROJECT

Your next idea
belongs here, too.

GPT-Vis grows with the people using it. Share a report, suggest a chart, or help build the next feature.

Explore on GitHub Share an idea

MIT licensed · Release notes ↗

antvis / GPT-Vis

GITHUB STARS794

A little star goes a long way.

96forks & counting