ChartGlow

Create a ChartGlow bar-race (default), line-race, bubble-race, or treemap-race project from a structured dataset prepared by ChatGPT using user-provided data or sources it has actually retrieved and verified. Include units, source references, time/entity/value mapping and visual configuration; values must be finite numbers, and bubble-race / treemap-race values must be positive. Bubble race renders positive values as relative bubble areas in an automatic force layout, normalized to the available space at each period and capped to stay inside the chart. Positions are not x/y data coordinates; compare the displayed numbers for absolute changes over time. The plugin uses native area sizing; radius sizing and bubbleScale are not exposed options. Treemap race uses native rectangular areas relative to each period total, with gutters and visual interpolation between observations; compare displayed values for absolute changes over time. It is not a bar ranking or a fixed dollar-area scale. Never invent statistics. ChartGlow validates and renders supplied data without fetching or generating it, calling Gemini or another AI, or independently verifying source contents. Supports at most 2,000 rows, 50 entities and 120 periods. Aspect ratio defaults to 9:16. options.visualTheme selects an existing native preset listed in options.visualTheme, including Infographic. For temporal trajectories prefer line-race, for rankings bar-race, and for within-period shares bubble-race or treemap-race when appropriate; preserve an explicit user choice. Never silently cumulate or transform data. options.decimals (0–6) and compact control native number display, dynamicBarHeight controls Bar geometry, and safeZone selects none/universal/tiktok/instagram/youtube/x; social safe zones require 9:16 or 4:5 and reuse the native layout. Select a theme only when requested; omitted visual options keep the native defaults. Keep country names as data; do not manufacture flag columns, country codes or image URLs. Country flags and custom images can be added with native controls in the ChartGlow editor through Edit in ChartGlow. Preserve supported assets already supplied with a dataset or an Explore example. Arbitrary external images and unsupported chart types are rejected. Generate directly with the available verified data and reasonable defaults; do not delay creation with nonessential preference questions. After creation, offer at most two optional ideas only for an identifiable need. After a small edit, briefly confirm the change; do not repeat suggestions, theme changes or MP4 instructions. Respect the user's language and existing choices. Use only supported settings and do not apply suggestions, change data, or start a transfer without the user choosing. When relevant to the request, explain MP4 once: the user clicks the widget export button, opens ChartGlow, signs in if needed and saves, then exports MP4 in the browser editor. Never claim that a project was saved or a video exported by create_chart or modify_chart.

create_chart

Read the two published Explore examples compatible with this plugin, without fetching arbitrary URLs, creating a chart, storing data or calling AI. Omit slug for two short choices with title, chart type, period, unit and links. After the user chooses, use apple-revenue (Treemap) or renewable-power (Bubble) to retrieve the COMPLETE dataset, mapping, supported options and provenance as selected.createChart, ready for create_chart. Also returns full source, rights and coverage notices and limits of the compatible composition; do not promise a pixel-exact Explore copy or other catalogue examples. Explain choices and relevant limitations in the user's language and respect explicit preferences. Never invent data, remove attribution, change units, cumulatively sum these annual/stock observations or claim a saved project/video. ChatGPT can still prepare verified structured data directly for create_chart without using an example.

get_explore_examples

Modify a complete ChartGlow bar-race, line-race, bubble-race, or treemap-race envelope supplied by the caller. Treemap race uses native rectangular areas relative to each period total, with gutters and visual interpolation between observations; compare displayed values for absolute changes over time. It is not a bar ranking or a fixed dollar-area scale. Can switch between these types with chartType, or replace dataset, mapping, or provenance when included in the patch; omitted type and data are preserved. Bubble race uses positive values for relative bubble areas normalized to available space at each period, with automatic force layout and no x/y scatter coordinates. Compare displayed numbers for absolute changes over time; native radius sizing and bubbleScale are not exposed plugin options. patch.visualTheme explicitly selects a native preset listed in its schema, including Infographic; an omitted theme preserves the current theme and customizations. A theme selection reapplies its native preset and palette; apply explicit colors afterward or in the same patch. Returns the full envelope with the same project id and revision incremented. ChatGPT prepares any replacement or edited dataset from user-provided data or sources it has actually retrieved and verified, with units and source references; never invent statistics. ChartGlow validates and renders supplied data without fetching or generating it or calling another AI. Preserve existing supported images when editing. Do not add flag columns, country codes or image URLs to country data; direct flag and image customization to the native ChartGlow editor through Edit in ChartGlow. Arbitrary external images are rejected. patch.decimals (0–6), compact and dynamicBarHeight expose native readability settings; patch.safeZone supports none/universal/tiktok/instagram/youtube/x (social presets require 9:16 or 4:5). Omitted settings are preserved; reapply a safe-zone preset explicitly after a ratio change when desired. This tool stores nothing. Generate directly with the available verified data and reasonable defaults; do not delay creation with nonessential preference questions. After creation, offer at most two optional ideas only for an identifiable need. After a small edit, briefly confirm the change; do not repeat suggestions, theme changes or MP4 instructions. Respect the user's language and existing choices. Use only supported settings and do not apply suggestions, change data, or start a transfer without the user choosing. When relevant to the request, explain MP4 once: the user clicks the widget export button, opens ChartGlow, signs in if needed and saves, then exports MP4 in the browser editor. Never claim that a project was saved or a video exported by create_chart or modify_chart.

modify_chart

After an explicit user click, prepare a short-lived transfer for the current chart and open its save/export page in the host. The complete current dataset, units, sources and configuration are preserved without regenerating data or calling AI. This creates temporary storage only; it does not save to an account or export a file.

prepare_save_export