@dotzero.ai/export-mcp

MCP server for chart generation and data export — PNG/JPG charts, CSV/XLSX files.

Overview

A pure rendering engine that generates charts and exports data files. No authentication required, no external API calls. Designed to work alongside other DotZero MCP servers: fetch data with existing tools, then visualize or export with this package.

Architecture: User → AI Agent → 1. Fetch data (existing MCP tools) → 2. Render/Export (export-mcp)

Installation

npx @dotzero.ai/setup

Or manually:

claude mcp add dotzero-export --command "npx" --args "@dotzero.ai/export-mcp"

Tools (14)

Generic Charts (5)

DotZero-Specific Charts (4)

Export (2)

Smart Tools (2)

Shared Parameters

All chart tools share these base parameters:

All export tools share:

Chart Options

Bar, line, and scatter charts accept an options object:

Tool Reference

chart_bar

chart_bar(
  title: "Production by Line",
  labels: ["Line A", "Line B", "Line C"],
  datasets: [
    { label: "Good", data: [1200, 980, 1150] },
    { label: "Defect", data: [23, 15, 31] }
  ],
  options: { stacked: true, y_label: "Quantity" }
)

chart_line

chart_line(
  title: "Weekly Production Trend",
  labels: ["Mon", "Tue", "Wed", "Thu", "Fri"],
  datasets: [
    { label: "Output", data: [450, 520, 480, 510, 530], fill: true }
  ],
  options: { y_label: "Units" }
)

chart_pie

chart_pie(
  title: "Defect Distribution",
  labels: ["Scratch", "Dent", "Crack", "Other"],
  values: [45, 28, 15, 12],
  chart_type: "doughnut"
)

chart_scatter

chart_scatter(
  title: "Cycle Time vs Output",
  datasets: [{
    label: "Devices",
    data: [
      { x: 30, y: 120 },
      { x: 45, y: 95 },
      { x: 25, y: 140 }
    ]
  }],
  options: { x_label: "Cycle Time (s)", y_label: "Output" }
)

chart_gauge

chart_gauge(
  value: 85.2,
  label: "OEE",
  min: 0,
  max: 100,
  thresholds: [
    { value: 60, color: "#F44336" },
    { value: 80, color: "#FFC107" },
    { value: 100, color: "#4CAF50" }
  ]
)

chart_oee_breakdown

Single device:

chart_oee_breakdown(
  title: "CNC-001 OEE",
  availability: 92.5,
  quality: 98.1,
  performance: 87.3
)

Multi-device comparison:

chart_oee_breakdown(
  title: "Line A OEE Comparison",
  availability: 0, quality: 0, performance: 0,
  devices: [
    { name: "CNC-001", availability: 92.5, quality: 98.1, performance: 87.3 },
    { name: "CNC-002", availability: 88.0, quality: 96.5, performance: 91.2 },
    { name: "CNC-003", availability: 95.1, quality: 97.8, performance: 84.6 }
  ]
)

chart_control

chart_control(
  title: "Diameter Control Chart",
  values: [10.02, 10.05, 9.98, 10.01, 10.12, 9.95, 10.03, 10.08],
  labels: ["S1", "S2", "S3", "S4", "S5", "S6", "S7", "S8"],
  ucl: 10.10,
  cl: 10.00,
  lcl: 9.90,
  usl: 10.15,
  lsl: 9.85
)

Points exceeding UCL/LCL are highlighted in red. USL/LSL specification limits are shown as orange dashed lines.

chart_timeline

chart_timeline(
  title: "CNC-001 State History",
  device_name: "CNC-001",
  segments: [
    { start: "2026-02-08T08:00:00Z", end: "2026-02-08T10:30:00Z", state: "running" },
    { start: "2026-02-08T10:30:00Z", end: "2026-02-08T11:00:00Z", state: "idle" },
    { start: "2026-02-08T11:00:00Z", end: "2026-02-08T11:45:00Z", state: "down" },
    { start: "2026-02-08T11:45:00Z", end: "2026-02-08T17:00:00Z", state: "running" }
  ]
)

Default state colors: running=green, idle=yellow, down=red, off=grey. Override with the color field per segment.

chart_multi

chart_multi(
  title: "Production Dashboard",
  layout: "grid",
  charts: [
    {
      type: "bar",
      title: "Output by Line",
      labels: ["Line A", "Line B"],
      datasets: [{ label: "Output", data: [1200, 980] }]
    },
    {
      type: "gauge",
      title: "OEE",
      value: 85.2,
      label: "OEE"
    },
    {
      type: "pie",
      title: "Defects",
      labels: ["Scratch", "Dent"],
      values: [65, 35]
    }
  ]
)

Layout options: grid (auto rows/cols), vertical (1 column), horizontal (1 row). Max 9 sub-charts.

export_csv

export_csv(
  headers: ["Work Order", "Product", "Qty", "Status"],
  rows: [
    ["WO-001", "Widget A", 100, "Completed"],
    ["WO-002", "Widget B", 250, "In Progress"],
    ["WO-003", "Widget A", 75, "Not Started"]
  ],
  filename: "work-orders"
)

export_xlsx

export_xlsx(
  sheets: [
    {
      name: "Work Orders",
      headers: ["ID", "Product", "Qty", "Status"],
      rows: [
        ["WO-001", "Widget A", 100, "Completed"],
        ["WO-002", "Widget B", 250, "In Progress"]
      ]
    },
    {
      name: "Summary",
      headers: ["Metric", "Value"],
      rows: [
        ["Total Orders", 2],
        ["Total Qty", 350]
      ]
    }
  ],
  filename: "production-report"
)

Headers are bold, column widths are auto-fit.

chart_from_json

chart_from_json(
  title: "Auto Chart",
  data: <raw JSON from any DotZero tool>,
  chart_type: "auto"
)

Auto-detection rules:

• Objects with availability/quality/performance fields → OEE breakdown chart
• Array of OEE objects → Multi-device OEE comparison
• Single number → Gauge chart
• Array of objects with name + numeric fields → Bar chart
• Force chart_type to override: "bar", "line", "pie", "gauge", "oee"

export_table_from_json

export_table_from_json(
  data: <raw JSON from any DotZero tool>,
  format: "xlsx",
  filename: "export",
  sheet_name: "Data"
)

Accepts arrays of objects, single objects, or paginated responses ({ data: [...] }). Automatically extracts column headers from object keys.

Workflow Examples

1. OEE Breakdown Chart (OEE 分解圖)

Step 1: oee_device(device_uuid: "...", start_time: "...", end_time: "...", response_format: "json")
Step 2: chart_oee_breakdown(title: "CNC-001 OEE", availability: 92.5, quality: 98.1, performance: 87.3)

Or use the smart tool:

Step 1: oee_devices(line_uuid: "...", response_format: "json")
Step 2: chart_from_json(title: "Line A OEE", data: <result from step 1>)

2. SPC Control Chart (SPC 管制圖)

Step 1: spc_measure_history_manufacture(config_uuid: "...", response_format: "json")
Step 2: spc_statistics_capability(config_uuid: "...", response_format: "json")
Step 3: chart_control(
  title: "Diameter Control",
  values: [extracted from step 1],
  ucl: [from step 2],
  cl: [from step 2],
  lcl: [from step 2]
)

3. Production Report Excel Export (生產報表匯出)

Step 1: workorder_list(status: 3, response_format: "json")
Step 2: export_table_from_json(data: <result>, format: "xlsx", filename: "completed-orders")

Or with custom formatting:

Step 1: production_summary(response_format: "json")
Step 2: export_xlsx(
  sheets: [
    { name: "Summary", headers: [...], rows: [...] },
    { name: "Details", headers: [...], rows: [...] }
  ],
  filename: "production-report"
)

4. Worker Efficiency Ranking (作業員效率排名)

Step 1: worker_efficiency_ranking(response_format: "json")
Step 2: chart_bar(
  title: "Worker Efficiency",
  labels: [worker names],
  datasets: [{ label: "Efficiency %", data: [values] }],
  options: { horizontal: true, y_label: "Worker" }
)

5. Device State Distribution (設備狀態分佈)

Step 1: equip_state_counts_factory(factory_uuid: "...", response_format: "json")
Step 2: chart_pie(
  title: "Device States",
  labels: ["Running", "Idle", "Down", "Off"],
  values: [running_count, idle_count, down_count, off_count],
  colors: ["#4CAF50", "#FFC107", "#F44336", "#9E9E9E"]
)

6. Multi-Chart Dashboard (多圖表儀表板)

Step 1: Fetch OEE, production summary, and defect data
Step 2: chart_multi(
  title: "Daily Dashboard",
  layout: "grid",
  charts: [
    { type: "gauge", title: "OEE", value: 85.2, label: "OEE" },
    { type: "bar", title: "Output", labels: [...], datasets: [...] },
    { type: "pie", title: "Defects", labels: [...], values: [...] },
    { type: "line", title: "Trend", labels: [...], datasets: [...] }
  ]
)

File Output

• Default path: .dotzero/exports/<type>-<name>-<timestamp>.<ext>
• .dotzero/ is in .gitignore — files won't be committed
• Override with output_path parameter on any tool

Output Format

Chart tools return:

{
  "content": [
    { "type": "text", "text": "Chart saved to: .dotzero/exports/chart-bar-2026-02-08T14-30-00.png (800x500)" },
    { "type": "image", "data": "<base64>", "mimeType": "image/png" }
  ]
}

Export tools return:

{
  "content": [
    { "type": "text", "text": "Exported 42 rows to: .dotzero/exports/export-2026-02-08.csv\n\nPreview (first 5 rows):\n| Col1 | Col2 |\n|---|---|\n| ... | ... |" }
  ]
}

DotZero Color Palette

Dependencies

Build

cd packages/export-mcp
pnpm install
pnpm build

License

MIT