Flourish Connector - an intro

Data sources and story ideas are often scattered in a multitude of places - which can turn creation into a lengthy affair. But time is a precious commodity, especially in the age of AI. Jumping between tools, transforming data for each one and trying things out, can keep the data storyteller focused on busy work instead of creation.

That’s where the Flourish Connector comes in. Using the Model Context Protocol (MCP) standard, you can connect any compatible AI tool, such as Claude.ai and ChatGPT to Flourish and pull together powerful capabilities, exactly where your data live. You can create and update charts through conversation with the AI tool of your choice, as long as it supports the MCP protocol.

Taking you from idea to polished project within a few clicks.

TIP: Read more about our AI offerings and the way we approach artificial intelligence in data storytelling, as a facilitator and not a replacement of creativity.

What is it for

The Flourish Connector is an MCP connector that lets compatible AI tools create Flourish visualizations from your data.  You provide or ask for a dataset, describe what you want to show, and the AI agent recommends the right chart type for your data, selects a template, and transforms the data to fit the template.

It then creates and configures the chart, and returns an edit URL, that you open in Flourish to polish and publish. The Connector can apply titles, subtitles, source lines, colours, fonts, and styling — so the chart is close to publish-ready before you ever open the Flourish editor.

Perfect for quickly trying out different approaches on a dataset, work with multiple data sources, iterate on an idea, or batch creation.

The Flourish connector works with Claude, ChatGPT, your own custom agent — anything that speaks MCP can connect to Flourish.

What you'll need before you start

A Flourish account, access to an AI tool that supports the MCP protocol, and a few minutes to configure the connection. If you don't have a Flourish account yet, you can create one here.

How to set up the Flourish Connector with Claude and ChatGPT

How to connect Flourish to ChatGPT

1

Head to Plugins, in the ChatGPT sidebar.

2

Search for Flourish.

Choose Sign in with Flourish.

3

Please take a moment to read our updated terms and privacy policy that accommodate our new AI features. If you are happy with them, proceed to Continue to Flourish.

Then - log in to Flourish.

4

The Connector should be working as expected and take you back to ChatGPT - so time to elevate your data storytelling. Please read our guide on how to explore chart options with AI and Flourish.

How to connect Flourish to Claude

1

Go to the sidebar on the left side of the screen and choose Customize. The UI is the same for Claude desktop and webpage.

2

Head to Connectors.

Click on Add and then Browse Connectors.

3

Search for Flourish - it is the connector you are looking for. Then click the plus sign.

Claude will then direct you to Flourish.

4

Please take a moment to read our updated terms and privacy policy that accommodate our new AI features. If you are happy with them, proceed to Continue to Flourish.

And then sign in to your Flourish account.

5

Once you log in to Flourish, you will be prompted to authorize Claude to access your Flourish account.

6

After a successful connection, you will be able to see Flourish in the list of connectors.

7

By clicking on the connector, you can control its permissions. 

However, as you will be using it and based on the different tasks the connector will have to complete, it will also be prompting you about its permissions.

 

How to connect Flourish to other AI tools

The Flourish Connector can be used with any AI tool that supports the MCP protocol, such as Cursor, Claude Code, Codex, Antigravity and others.

Claude Code

To connect the Flourish Connector to Claude Code using the MCP protocol, run the following command in your terminal:

claude mcp add --transport http flourish https://app.flourish.studio/mcp

Then  run /mcp in Claude Code to authenticate by following the OAuth flow.

Cursor

Open Cursor Settings, then head to MCP and choose Add new global MCP server.

Paste the following

{
  "mcpServers": {
    "flourish": {
      "url": "https://app.flourish.studio/mcp"
    }
  }
}

Codex

To add the Flourish Connector to Codex, you must add the Flourish server to your Codex configuration at ~/.codex/config.toml:

[mcp_servers.flourish]
url = "https://app.flourish.studio/mcp"

And then run the following to begin the authentication flow:

codex mcp login flourish

Gemini

If using the Gemini CLI, run the following:

gemini mcp add --transport http flourish https://app.flourish.studio/mcp

Alternatively, you can add the server to ~/.gemini/settings.json under mcpServers.

Other tools

If your AI tool isn’t listed above but supports MCP, you can connect the following URL and specifications:

Server URL: https://app.flourish.studio/mcp
Connection type: OAuth
Transport: Streamable HTTP

How to use the Flourish MCP Connector with Claude

1

Head back to the Claude interface you are using - chat, Cowork, or terminal.

TIP: Searching for ideas? Read our guide on how to make the ideal popup in 4 prompts.

2

Upload the dataset you want to use, paste in a link to it, or just ask Claude to provide a dataset. 
If you opt to upload data directly — the dataset can be up to 1,000 rows or 1MB. For larger datasets, the agent will guide you on alternatives.

Moreover, as MCP is an open protocol, your AI tool can pull data from other connected sources — Google Sheets, databases, APIs — and pass it straight into Flourish.

Warning: If the agent needs to reshape your data — for example, from long to wide format —, it can occasionally produce incorrect values, especially with large or complex datasets. Please double-check your data.

3

Claude will begin chatting with you about the various visualization options, depending on your original ask, and also ask for the necessary permissions to do its work in Flourish.

TIP: If the agent infers a column's type incorrectly — for example, treating dates as text — it can't fix that programmatically. You must adjust data types in the Flourish editor.

4

The AI agent will provide an edit link to the Flourish visualization it created.

Please note that the MCP respects your existing Flourish permissions. If you have edit access to a visualisation, the agent can make changes to it — useful for quick styling tweaks without opening the editor.

5

To fine-tune things, like annotation placement, axis formatting, and legend positioning, click the edit link and head to the Flourish Editor, newly empowered with the Flourish Assistant.

FAQ/Troubleshooting

Q: I can't seem to be able to use the Flourish Connector?

If you are part of a Company account, please make sure that your company admin has clicked Enable Connector in the company settings, and also that your individual account has the feature enabled.

For personal accounts, head to your account settings and make sure you have enabled AI-powered features.

Q: Will the MCP use up my Flourish AI credits?

For the Flourish Connector, token cost sit entirely with your chosen AI tool and not with Flourish.

Q: Can multiple team members use the same MCP connection?

No, the MCP is authenticated on a single user level.

Q: Will my company branding be upheld - can I call specific themes and have them applied to MCP-created visuals? 

If a default theme exists, it will be applied to newly created visualizations, similar to the editor experience. The MCP can't change or switch themes.