bulkrender-mcp

BulkRender MCP server — generate DOCX and PDF documents from any AI assistant that supports the Model Context Protocol (Claude, Cursor, Windsurf, Cline, and more).

Tools

Tool	Description	Credits
`list_templates`	List all templates with their variables	0
`get_template`	Get template details and variable schema	0
`generate_document`	Generate a single document (DOCX or PDF)	1–2
`generate_batch`	Generate documents for multiple records (up to 500)	1–2 each
`check_credits`	Check remaining credits	0

Prerequisites

A BulkRender account at app.bulkrender.com
An API key (br_live_...) — generate one from Settings → API Keys
Node.js 18+

Installation

Claude Desktop / Claude Code

Add to your Claude MCP config (~/.claude.json or Claude Desktop settings):

{
  "mcpServers": {
    "bulkrender": {
      "command": "npx",
      "args": ["-y", "bulkrender-mcp"],
      "env": {
        "BULKRENDER_API_URL": "https://api.bulkrender.com"
      }
    }
  }
}

Cursor / Windsurf / Cline

Use the same npx -y bulkrender-mcp command pattern in your MCP client settings.

Configuration

Variable	Default	Description
`BULKRENDER_API_URL`	`https://api.bulkrender.com`	BulkRender API base URL

Set BULKRENDER_API_URL=http://localhost:3000 for local development.

Usage Examples

List templates:

"List my BulkRender templates"

Generate a document:

"Generate an invoice using the Invoice Template for Acme Corp, invoice #1234, dated 2025-01-15, amount $5,000"

Batch generation:

"Generate invoices for these 3 clients: Acme Corp ($5,000), Beta Inc ($3,200), Gamma LLC ($7,800)"

Check credits:

"How many BulkRender credits do I have left?"

Rate Limits

All limits are enforced by the BulkRender API, not the MCP server itself.

Endpoint type	Limit
Document generation (`generate_document`, `generate_batch`)	30 requests / minute per organization
Template reads (`list_templates`, `get_template`, `check_credits`)	30 requests / minute per organization

When the limit is hit, the API returns HTTP 429. The MCP server surfaces this as a tool error with the message "You are making too many requests. Please try again later." — the AI assistant will see this and can retry after a short pause.

Batch jobs: generate_batch counts as one API request regardless of record count (up to 500 records). Use batching to stay well within the rate limit when generating many documents at once.

Security

API key is passed per-tool-call, never stored by the MCP server
All production communication uses HTTPS
Download URLs are signed and expire after 1 hour

bulkrender

Quick Install