Guides

From zero to your first PDF.

• Quickstart — your first PDF in 5 minutes

  Sign up, design a template, render your first PDF through the API. End-to-end in five minutes.

Connect Claude, Cursor, ChatGPT, or Gemini to your templates.

• Use Transactional from your AI assistant

  Connect Transactional to Claude, Cursor, ChatGPT, or Gemini via MCP so your assistant can read your templates and generate PDFs directly.

• Claude Desktop

  Connect Transactional to Claude Desktop on macOS or Windows so Claude can read your templates and generate PDFs.

• Claude Code (CLI)

  Connect Transactional to Claude Code so it can read your templates and generate PDFs from your terminal.

• Cursor

  Wire Transactional into Cursor's MCP support so you can generate PDFs from inside your editor.

• ChatGPT (via Custom GPT Actions)

  ChatGPT doesn't speak MCP natively yet — but you can give it the same powers through a Custom GPT pointed at the Transactional REST API.

• Gemini Code Assist / Gemini CLI

  Connect Transactional to Google's Gemini agents through their MCP support.

• MCP tools reference

  Every tool the Transactional MCP server exposes, with arguments, return shapes, and a prompt that typically triggers each one.

Call /v1/generate from your stack — Node, PHP, Python.

• Calling /v1/generate from Node.js & Bun

  Production-grade integration using native fetch — retries, error handling, streaming the PDF to your storage.

• Calling /v1/generate from PHP & Laravel

  cURL extension, Guzzle, or Laravel's HTTP client — render PDFs with retries and proper error handling.

• Calling /v1/generate from Python

  urllib (stdlib), requests, or httpx with retry — render PDFs from Django, FastAPI, or any Python service.

Handlebars, design for print, variable modeling, AI editor.

• Handlebars cheat sheet

  The exact subset of Handlebars supported in Transactional templates — variables, conditionals, loops, and what NOT to reach for.

• Designing templates that survive PDF rendering

  PDFs are static — drop the animations, oversample your canvas charts, lean on vectors. The rules that make a template look sharp at print resolution.

• Modeling your variables

  When to make something a variable vs. inline. Keep the API contract small, your templates portable, and your integration code boring.

• Working with the AI assistant

  Prompts and patterns to get good templates fast — what to ask, when to iterate, when to start over.

Monitor usage, store and serve PDFs at scale.

• Monitoring usage & credits

  Read the dashboard gauges, set sane alerts, and decide when to top up vs. upgrade.

• Storing & serving the generated PDF

  The /v1/generate URL is signed and short-lived. Patterns for keeping the PDF around — your bucket, your CDN, your DB.