datawrapper-mcp: Datawrapper charts through an agent
datawrapper-mcp
An MCP server and app for Datawrapper: an agent creates, updates, publishes and exports charts to PNG from your data
- 64 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The token opens creating, publishing and deleting charts in the Datawrapper account
- A published chart is available by public link and carries data from the chat
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/datawrapper-mcpDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The token opens creating, publishing and deleting charts in the Datawrapper account
- A published chart is available by public link and carries data from the chat
This is third-party code. Review the repository files before installing.
What it does
The server is built on the datawrapper Python library and gives the agent eight tools. list_chart_types and get_chart_schema describe the available chart types and the settings schema of a type. create_chart creates a chart from data and configuration, update_chart changes data or styling, publish_chart makes a chart available by a public link, get_chart returns settings and metadata, export_chart_png returns an image and delete_chart removes a chart permanently. Supported types are bar, line, area, arrow, column, multiple column, scatter and stacked bar. A data file can go straight into chat, and the draft appears in the app panel. After publishing the agent returns the public URL and the editor URL, and on a request for a PNG it embeds the image in the reply. When connecting to a hosted server over HTTP you can pass your token in the Authorization header or as the access_token argument so charts are created in your own account. The DATAWRAPPER_MCP_INSTRUCTIONS variable adds your own house style rules, for example mandatory alt text.
Who it is for. Journalists, analysts and data desks who build charts in Datawrapper and want drafts from a table in one sentence.
Good fit when
- You need a quick line chart from a CSV table and a tweak to the line color
- A published chart needs a new data point for the latest year
- An article needs a PNG version of a chart that the agent will embed in the reply
- A newsroom wants unified rules: alt text and captions set through DATAWRAPPER_MCP_INSTRUCTIONS
- The agent should suggest five ways to improve an existing chart
Not a fit when
- You need Datawrapper maps and tables: the server supports only eight chart types
- You have no Datawrapper account: the server does not work without an API token
- You want installation through a plugin directory without editing config: that path is still experimental
Example request
Build a line chart of temperatures from this data, publish it, add a point for 2024 and show the PNG.MCP
- Transport
- stdio, http
- Authentication
- API key
Environment variables
DATAWRAPPER_ACCESS_TOKENrequired, secret- Datawrapper API token
DATAWRAPPER_MCP_INSTRUCTIONS- Custom house style guidance for clients
Limitations
The project is early and experimental and behavior may change between releases. Installation through plugin directories, including the Claude Desktop plugin, cannot yet collect environment variables such as DATAWRAPPER_ACCESS_TOKEN, so the reliable path is one: editing config by hand. You need a Datawrapper account, an API token from app.datawrapper.de/account/api-tokens and Python 3.10 or newer. delete_chart removes a chart permanently, and a published chart is open by public link. Detailed steps for Claude Desktop, Cursor, Codex and Kubernetes are in INSTALLATION.md.
How to disable. Remove the datawrapper entry from the client config and revoke the token on the Datawrapper api-tokens page. Stop and remove the container from the Docker image.
FAQ
Which chart types are available?
Bar, line, area, arrow, column, multiple column, scatter and stacked bar. list_chart_types gives descriptions and get_chart_schema gives the full settings schema for a type.
Where do I get a token?
Create a token on app.datawrapper.de/account/api-tokens and put it in the DATAWRAPPER_ACCESS_TOKEN variable in the MCP config.
Which clients fit?
Claude Desktop, Claude.ai, Claude Code, Cursor, VS Code Copilot, ChatGPT, OpenAI Codex and OpenClaw. The transport depends on the client, stdio or streamable-http.
How do I use a hosted server under my own account?
Send the token in the Authorization: Bearer header or as the access_token argument. The argument takes precedence over the header, and without both the server falls back to its environment variable.
Is there a Docker image?
Yes, the palewire/datawrapper-mcp image is on Docker Hub, and the installation guide covers Kubernetes.
Discussion
No questions or reviews yet.
Related
MCP serversMCP Dashboards: interactive charts in chat
MCP Dashboards
32 tools draw interactive charts, KPI widgets and dashboards right inside the chat in Claude Desktop, VS Code and other clients
An MCP server from AntV for generating charts and diagrams: 25+ visualization types, from lines and bars to maps, mind maps and network graphs
Official Data Commons MCP server: an agent finds statistical indicators and pulls observations for countries, regions and cities.
Official Tableau MCP server: the agent finds data sources, sends queries to them and opens dashboards and Pulse metrics
MCP server and skills for product analytics: user paths, funnels, segments and transition graphs over event logs
MCP server and CLI that turns paper text or data into academic figures, diagrams and plots
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| datawrapper-mcp: Datawrapper charts through an agent (this card) | no | API key | 64 | |
| microcharts MCP | no | not required | 208 | compact microcharts for embedding, without publishing to Datawrapper |
| AntV Chart Visualization Skills | no | 535 | chart-building skills, no Datawrapper server connection |
README in short
The README warns about the early stage, shows a chart creation and publishing dialogue, a table of eight tools, chart types and requirements. Token setup, hosted server use, custom instructions and a supported client table follow.