← Back to directory
B

Bilig WorkPaper

Community
Run workbook-shaped business rules inside Node with formula readback.
GitHub source repository ↗
★ 36 Stars Category · Dev Tools Popular Source revision 3f740671e846
57FMRS · C

The Bilig WorkPaper MCP server brings spreadsheet formula calculations to AI agents and Node services, supporting input edits, formula recalculation, JSON persistence, and an MCP tool interface. It is integrated with the official MCP registry and provides both stdio and remote HTTP endpoints. Although not officially maintained, it offers thorough verification paths in its documentation.

Reliability
10/20
Security and permissions
13/20
Maintenance
14/20
Documentation
10/20
Setup experience
10/20
Read the FMRS scoring method →

Bilig WorkPaper is an MCP server that provides agents formula-workbook read/write and recalculation capabilities. It enables editing input cells, reading formula-computed outputs, persisting workbook state as JSON, and exposing MCP tools for workbook risk preflight. Designed for Node services and AI agents that need to interact with structured formula workbooks.

Tools

read_range
Read cell values from a specified range.
read_cell
Read the value of a single cell.
set_cell_contents
Set the contents of a cell and trigger formula recalculation.
set_cell_contents_and_readback
Set cell contents and read back the dependent formula result.
get_cell_display_value
Get the display value of a cell.
export_workpaper_document
Export the workbook as a JSON document.
validate_formula
Validate the syntax and correctness of a formula.
analyze_workbook_risk
Analyze workbook import/export risks.

Setup

Launch the server via stdio using the npm package @bilig/headless, or use the remote Streamable HTTP endpoint https://bilig.proompteng.ai/mcp. To configure an MCP client, add the JSON above to your client config file.

claude_desktop_config.json
{
  "mcpServers": {
    "bilig-workpaper": {
      "command": "npx",
      "args": [
        "-y",
        "@bilig/headless"
      ]
    }
  }
}

Fit and risk

Best for

  • Scenarios requiring exact cell addresses with computed readback
  • Formula-driven business logic in backend services
  • Workbook applications needing JSON persistence

Not for

  • Visual spreadsheet editing
  • Office macros or desktop Excel automation
  • One-off arithmetic calculations

Required permissions

  • Read and write workbook cells
  • Export JSON documents locally or remotely

Risks and side effects

  • Limited formula compatibility; not all Excel functions supported
  • Embedded XLSX cached values may be inaccurate; use Excel recalculation as oracle

Troubleshooting

  1. Run bilig-evaluate --door agent-mcp to verify MCP connectivity
  2. Check Node version >=22
  3. Ensure workbook file path is correct

Use cases

AI agents editing spreadsheet inputs and reading calculated outputs
Pricing models, quote approvals, payout checks in Node services
Verifying formula correctness in CI/CD tests
Saving and restoring workbook state in workflow automation

Supported clients

Claude DesktopFull support
OpenAI Agents SDKFull support
n8nFull support

More servers from this repository