How to build a report with tables and charts from a CSV or spreadsheet with an AI agent Leave feedback

Run the GroupDocs.Assembly MCP server locally, point your agent at a spreadsheet or CSV, and it produces the report your template defines — the rows, the total, the best performer, and a chart — without drawing a single table itself:

GROUPDOCS_MCP_STORAGE_PATH=$(pwd)/templates dnx GroupDocs.Assembly.Mcp --yes

Build the Q3 sales report from sales.xlsx as a PDF, and the chart slide from the same data.

Note

The commands and config snippets on this page are for the .NET build of the server — the only platform available today. Installation and client setup: MCP server for .NET. Other platforms will expose the same tools with their own launch command; everything else on this page applies unchanged.

• The server registered in your AI client — quick start.
• A report template with a table row that repeats, and — for a chart — a PowerPoint, Word or Excel template whose chart carries the tags. On Linux with dnx, presentation templates and charts need libgdiplus (Linux install); the Docker image has it.

sales.xlsx (table 1: Region, Quarter, Revenue)
     |
     +-- assemble_document  template sales-report.docx  --> table rows,
     |                                                      total, best region
     '-- assemble_document  template sales-chart.pptx   --> column chart

sales-chart.pptx, as designed in PowerPoint:
  chart title   Revenue by region<<foreach [r in sales]>><<x [r.Region]>>
  series name   Revenue<<y [r.Revenue]>>
after assembly:
  categories    North    South    East    West
  Revenue       182000   146500   98250   211900

The table template:

row 1 (header):  Region | Quarter | Revenue (EUR)
row 2 (repeats): <<foreach [r in sales]>><<[r.Region]>> | <<[r.Quarter]>>
                 | <<[r.Revenue]>><</foreach>>
Total revenue: <<[sales.Sum(r => r.Revenue)]>> EUR
Best region: <<[sales.OrderByDescending(r => r.Revenue).First().Region]>>

Captured from server 26.11.0 (the evaluation note that opens each response is trimmed):

-> get_document_info {"file":{"filePath":"sales.xlsx"}}
<- { "dataKind": "table", "data": { "tableCount": 1, "tables": [
     { "number": 1, "columns": [ "Region", "Quarter", "Revenue" ] } ] } }

-> assemble_document {"template":{"filePath":"sales-report.docx"},
     "data":{"filePath":"sales.xlsx"},"format":"pdf"}
<- Assembled 'sales-report.docx' with 'sales.xlsx' (table 1 (3 column(s):
   Region, Quarter, Revenue) as 'sales') as PDF
   Saved to: …/sales-report_assembled.pdf

-> assemble_document {"template":{"filePath":"sales-chart.pptx"},
     "data":{"filePath":"sales.xlsx"}}
<- Assembled 'sales-chart.pptx' with 'sales.xlsx' (table 1 (3 column(s):
   Region, Quarter, Revenue) as 'sales') as PPTX
   Saved to: …/sales-chart_assembled.pptx

-> assemble_document {"template":{"filePath":"sales-report.docx"},
     "data":{"filePath":"sales.csv"},
     "outputName":"sales-report-from-csv.docx"}
<- Assembled 'sales-report.docx' with 'sales.csv' (CSV with a header row
   as 'sales') as DOCX

The report lists North, South, East and West with their revenue, Total revenue: 638650 EUR and Best region: West. The chart came out with the four regions as categories and 182000, 146500, 98250 and 211900 as the series. The same template took the CSV unchanged.

Design the chart in PowerPoint (or Word, or Excel) with placeholder data, then put the tags in the chart’s text, not in a table:

• the chart title opens the loop and names the category: Revenue by region<<foreach [r in sales]>><<x [r.Region]>>;
• each series name names its value: Revenue<<y [r.Revenue]>>.

The engine replaces the placeholder data with one point per record and removes the tags. More chart types: the library’s template syntax.

A PDF or Word report and a chart slide whose numbers match the data file, in seconds. Licensed, no evaluation line.

Can the model draw the chart instead? It can describe one; it cannot put a native, editable chart in your deck. The template’s chart is a real PowerPoint/Word/Excel chart that people can restyle.

Which table does it read from a workbook with several sheets? Table 1 by default; dataTable: 2 picks the second. get_document_info lists them.

Can the output be Excel? Yes — an XLSX template with tags in its cells. Unlicensed, it gains an evaluation worksheet.

Agents follow project instructions about when to use a tool. Paste this into the instruction file your agent reads (CLAUDE.md, AGENTS.md, .github/copilot-instructions.md, .cursorrules):

For reports from ./data/*.xlsx or *.csv, call get_document_info on the data,
then assemble_document with the report template in ./templates and data as a
file. Never compute or retype the numbers yourself; quote totals from the output.

• Evaluation mode: every generated document starts with an Evaluation Only line, workbooks gain a worksheet, and e-mail templates lose their HTML body. Licensing.
• <<[name]>>, not {{name}}: other engines’ placeholders stay in the output as text.
• Templates are trusted input: image and document tags load the paths and URLs their expressions return.
• Dates: write them as "2026-11-12T00:00:00" and format them in the template; a date-only string can shift by a day on machines east of UTC.

• Generate documents with AI agents
• One document per record (mail merge)
• Inspect and fix a template
• From notes or email to a document
• Reports from repository data
• On-premise deployment
• Tools reference · Licensing · Troubleshooting & FAQ
• Questions: Assembly forum

• The chart comes from the template, the numbers come from the data — the same scenario on the GroupDocs blog, with worked examples and the trade-offs.