Last reviewed
May 9, 2026
Sources
13 citations
Review status
Source-backed
Revision
v2 ยท 2,498 words
Add missing citations, update stale details, or suggest a clearer explanation.
Data visualization ChatGPT plugins were a group of third party tools that extended ChatGPT so that the chatbot could draw charts, plot graphs, render flowcharts, and build other visual representations of data inside the chat interface. They were one category among many in the broader ChatGPT Plugins ecosystem, which OpenAI opened in beta on March 23, 2023, expanded to all ChatGPT Plus subscribers on May 12, 2023, and then wound down between March 19, 2024 and April 9, 2024.[1][2][3] During that roughly twelve month window, plugins such as Show Me Diagrams, daigr.am, the Wolfram plugin, QuickChart, and Noteable gave the chatbot a way to turn natural language requests into pictures without leaving the conversation.
The data visualization category sat next to closely related categories such as Data Analysis ChatGPT Plugins, research tools, productivity tools, and developer tools. Where data analysis plugins focused on loading, transforming, and querying data, visualization plugins concentrated on the final rendering step: producing an image, an interactive chart, or a diagram that a user could view, edit, share, or paste into a document. In practice the two categories overlapped heavily, and several plugins lived in both worlds at once.
When OpenAI introduced ChatGPT in late 2022, the product was a pure text interface. The model could describe a chart in words or write code that drew a chart elsewhere, but it could not render images directly in the conversation. The plugin system, announced on March 23, 2023, changed this by giving the model a way to call external services and embed their responses in a reply.[1] OpenAI launched the system with a small set of partners, including Expedia, FiscalNote, Instacart, KAYAK, Klarna, Milo, OpenTable, Shopify, Slack, Speak, Wolfram, and Zapier.[1] Wolfram was the only launch partner whose primary value proposition included visualization.
A broader rollout followed in waves. On May 12, 2023, OpenAI made the plugin store available to ChatGPT Plus subscribers, and the catalog grew quickly past two hundred entries.[3] Visualization plugins arrived during this expansion. Noteable announced its plugin on May 11, 2023.[4] Show Me Diagrams was added in late June 2023.[5] The daigr.am plugin and several smaller chart utilities followed over the summer.
The ecosystem peaked in the second half of 2023. After OpenAI introduced Custom GPTs on November 6, 2023 and the GPT Store on January 10, 2024, the company shifted developer attention to the GPT framework.[6][7] On the same day as the store launch, OpenAI confirmed that the plugin store would close. The official sunset followed in two stages: on March 19, 2024 users could no longer install new plugins or start fresh plugin conversations, and on April 9, 2024 all remaining plugin chats stopped working.[2]
The table below summarises the ecosystem timeline that framed every data visualization plugin.
| Date | Event |
|---|---|
| March 23, 2023 | OpenAI announces ChatGPT plugins in alpha with a dozen launch partners[1] |
| May 11, 2023 | Noteable announces its ChatGPT plugin[4] |
| May 12, 2023 | Plugin store opens to ChatGPT Plus subscribers[3] |
| June 2023 | Show Me Diagrams listed in the plugin store[5] |
| July 6, 2023 | Code Interpreter rolls out in beta to Plus users, providing native chart rendering[8] |
| September 2023 | Code Interpreter is renamed Advanced Data Analysis[8] |
| November 6, 2023 | OpenAI introduces Custom GPTs at DevDay[6] |
| January 10, 2024 | GPT Store opens; plugin sunset announced[7] |
| March 19, 2024 | Plugin installs and new plugin chats disabled[2] |
| April 9, 2024 | All plugin conversations end[2] |
A ChatGPT plugin was a small web service that exposed an OpenAPI specification, a manifest file, and an icon. When a user enabled a plugin, ChatGPT learned which actions the plugin offered and could call those actions during a conversation if it judged them helpful. The model formed the request, the plugin returned data, and ChatGPT folded the response back into its reply.
For the data visualization category, the response usually contained either an image URL or a text payload the chat interface could render as a diagram. Three rendering patterns dominated:
Because the system relied on the model deciding when to call a tool, prompt phrasing mattered. Skill at writing these requests overlapped with general prompt engineering practice for the underlying GPT-4 model.
The following plugins are documented in primary sources from the period and were widely covered while the plugin store was active. The list is not exhaustive. Hundreds of small plugins came and went without leaving public records, and many have been omitted because their existence cannot be reliably confirmed.
| Plugin | Primary output | Notes |
|---|---|---|
| Show Me Diagrams | Mermaid and PlantUML diagrams rendered as images | Built by an independent developer; published in June 2023; commonly cited as the most popular diagramming plugin[5][9] |
| daigr.am | Bar, line, pie, and donut charts | Plugin oriented toward simple business charts; required GPT-4 and an active plugin session[10] |
| Wolfram | Mathematical plots, statistical charts, geospatial maps, function graphs | Released March 23, 2023 as a launch partner; part of OpenAI's first plugin announcement; backed by Wolfram Language and Wolfram Alpha[1][11] |
| QuickChart | Static chart images and QR codes | Wrapper around the open source QuickChart service, which in turn uses Chart.js[12] |
| Noteable | Charts and notebooks rendered inside Noteable's Jupyter style workspace | Announced May 11, 2023; combined data analysis with visualization in a hosted notebook[4] |
| Mermaid Chart | Mermaid diagrams created and saved into the Mermaid Chart editor | Built by the team behind the Mermaid open source project; later reissued as a Custom GPT after the plugin sunset[13] |
Several other names appeared in third party listicles from 2023, including Diagrams: Show Me (a name some directories used for Show Me Diagrams) and assorted chart wrappers. Where a plugin could not be cross checked against the developer's own announcement or coverage in a reputable outlet, it has been left out.
Show Me Diagrams was the most visible plugin in the visualization category. It accepted natural language descriptions and produced diagrams in Mermaid syntax, which the plugin server then rendered as an image and embedded in the chat reply. Supported diagram families included flowcharts, sequence diagrams, mind maps, pie charts, timelines, class diagrams, entity relationship diagrams, and Gantt style timelines.[5][9] The plugin also supported PlantUML for diagram styles that Mermaid could not yet draw. Before the plugin existed, a user could ask ChatGPT to write Mermaid code, but they then had to copy that code into another tool to see the picture. The plugin removed that round trip.
The Wolfram plugin shipped on the first day of the plugin program and remained one of the most capable visualization tools throughout the program's life.[1][11] Because it called Wolfram Language and Wolfram Alpha, it could plot mathematical functions, generate statistical charts, draw three dimensional surfaces, render geographic maps, and produce specialised scientific diagrams that no other plugin attempted. Stephen Wolfram described the integration as giving ChatGPT "computational superpowers," and the plugin's visualization output was effectively whatever the underlying Wolfram engines could draw.[11]
The daigr.am plugin focused on the business chart use case: bar charts, line charts, pie charts, and donut charts assembled from values mentioned in the conversation. It required GPT-4 and was invoked by asking ChatGPT to "chart" or "graph" some figures.[10] QuickChart wrapped the open source QuickChart service, a long running tool that turns a JSON chart specification into a PNG image using the Chart.js library, and was often used for highly customised business charts.[12]
Noteable's plugin sat at the boundary between visualization and data analysis. Announced on May 11, 2023, it linked ChatGPT to a hosted Jupyter style notebook environment.[4] Users could ask ChatGPT to load a dataset, clean it, run a calculation, and then plot the result, all of which happened in the Noteable workspace and were mirrored back into the chat. The Mermaid Chart plugin, published by the company behind the commercial Mermaid Chart editor, produced Mermaid syntax in response to a natural language description and offered a link to open and refine the diagram in the Mermaid Chart web editor.[13] After the plugin sunset, the team rebuilt the same workflow as a Custom GPT, illustrating the wider migration path from plugins to GPTs.
The most common workflow began with a user asking ChatGPT for a picture of something. Examples included "draw a flowchart of the user signup process," "plot the sales numbers from the table above as a bar chart," or "show me a sequence diagram for an OAuth login." When a visualization plugin was enabled, ChatGPT would interpret the request, decide which plugin to call, build a structured payload, send it to the plugin server, and embed the returned image or link in its reply.
A second workflow used the chatbot to refine an output. After a plugin returned a chart, the user might say "change the bars to horizontal," and ChatGPT would call the plugin again with adjusted parameters. Plugins that stored work outside the chat, such as Mermaid Chart and Noteable, could persist iterations more cleanly. A third workflow used the chatbot as a translation layer: users who knew what they wanted in business terms but did not know Mermaid syntax, Chart.js JSON, or Wolfram Language could describe the desired output in English and rely on ChatGPT to produce the right input for the plugin.
OpenAI gave several reasons for deprecating the plugin store.[2][6][7] First, the plugin model required the chatbot to choose a single plugin per session in many cases, which limited how much could be combined in one conversation. Custom GPTs let a developer ship a packaged assistant with instructions, knowledge files, and multiple actions baked in. Second, plugin discovery was awkward: users had to enable plugins manually before each chat, and the plugin store interface was a separate panel rather than a first class part of ChatGPT. The GPT Store integrated discovery directly into the product. Third, the underlying chat experience improved on its own. Native Code Interpreter rendering, image generation, and later browsing made many plugin functions redundant.
For the visualization category, the third factor mattered most. By late 2023 the standard ChatGPT interface, running on top of GPT-4, could already produce charts and plots through Code Interpreter without any plugin at all. Many of the use cases that drove plugin installs in spring 2023 had been absorbed into the base product within six months.
After the plugin sunset, the workflows that had relied on visualization plugins moved to three main destinations.
Code Interpreter and Advanced Data Analysis. OpenAI launched Code Interpreter in beta on July 6, 2023, then renamed the feature Advanced Data Analysis in September 2023.[8] In 2024 the same feature was rebranded again as part of the standard ChatGPT toolset. Code Interpreter ran user supplied data through a sandboxed Python environment with libraries such as Matplotlib, Seaborn, and Plotly, and rendered the resulting charts directly in the chat. For most chart and plot needs, this internal feature replaced both data analysis and data visualization plugins.
Custom GPTs. Many plugin developers rebuilt their tools as Custom GPTs.[6][13] A Custom GPT could include the same OpenAPI actions a plugin used, plus instructions, conversation starters, and uploaded files. Mermaid Chart, Show Me, daigr.am, Noteable, and others all reappeared in some form in the GPT Store within months of the plugin shutdown. The GPT version of each tool was usually closer to a packaged app than its plugin predecessor.
Native ChatGPT chart rendering. During 2024 OpenAI added a native chart renderer to ChatGPT itself, so that simple bar, line, and pie charts could be drawn without invoking any tool. Combined with image generation and Code Interpreter, this reduced the appetite for stand alone visualization plugins still further.
The data visualization category was always intertwined with the data analysis ChatGPT plugins category. Plugins such as Noteable and Wolfram appeared in both directories, because the same underlying engine could load data, run a calculation, and then draw a picture of the result. The conventional split treated visualization plugins as the ones whose primary user value was a chart or diagram, and data analysis plugins as the ones whose primary value was a number, a transformed dataset, or a statistical answer. In practice users rarely cared about the boundary.
The convergence accelerated after the plugin shutdown. Code Interpreter united analysis and visualization in one feature inside ChatGPT, and Custom GPTs allowed developers to combine them in one packaged assistant. The categorical distinction that had structured the original plugin store became largely historical.
For about a year, data visualization plugins offered the first taste of what a chatbot looked like when it could draw. They proved natural language was a workable interface for chart making. Even though the plugin platform was deprecated, the workflows established during its lifetime carried forward into Code Interpreter, Custom GPTs, and the rendering features built into ChatGPT. Many of the original plugin developers continue to ship the same kinds of tools today inside the GPT Store.
See also: ChatGPT Plugins, ChatGPT Plugin Categories and Data Visualization