# What We Shipped Since the Summer


Hey everyone!

We have a small confession to make. Over the last few months, we’ve shipped quite a lot… and **told you about very little of it**. We found out the hard way: some of you wrote to us asking for features that were **already live**. That’s on us!

So here’s a quick catch-up on everything that landed since the summer. 🍂

## TL;DR

- **Rotate your API key** without downtime: the previous one keeps working for 2 days.
- **Snippet folders** to keep your snippets tidy, in the dashboard and through the API.
- **Fixed IP addresses** on engine v5, so you can allowlist PDFMonkey on your servers.
- Our **Zapier** app no longer races against Zapier’s 30-second limit.
- Our **n8n node** reaches 1.0, sets up its own webhook, and is **verified by n8n**.
- An official **Node.js SDK**, and an **MCP server** for Claude, Cursor and friends.

## Rotate your API key, without the panic

Changing a secret used to be a stressful moment: generate a new key, then rush to update every app still using the old one before something breaks.

Now, on the **API Key** page of your dashboard, the **Rotate API key** button gives you a new key while **your previous key keeps working for 2 days**. Plenty of time to update your integrations calmly. Done early? Click **Revoke previous key now** and the old key stops working right away. Otherwise, it expires on its own.

One rotation at a time, though: you can rotate again once the previous key has expired or been revoked. More on how your key is used in [API authentication](https://pdfmonkey.io/docs/api/authentication/).

## Folders for your snippets

Templates have had folders for a while. **Snippets now get them too.** Create folders from the **Snippets** page, move snippets into them with the **Move** action, and search across every folder at once.

The [Snippets API](https://pdfmonkey.io/docs/api/snippets/) is now fully documented as well, including [snippet folders](https://pdfmonkey.io/docs/api/snippets/#snippet-folders) and a `q[folders]` filter to list the snippets of a given folder.

## Fixed IP addresses for your resources

When we generate a document, our engine loads everything your template references: images, fonts, stylesheets, and any API your JavaScript calls. Until now, those requests came from **cloud IP addresses that kept changing**. Not great when your server only accepts known addresses, or when a CDN or firewall decides to block AWS addresses altogether.

Templates on **engine v5** now fetch their resources from **fixed IP addresses**, so you can allowlist PDFMonkey on your servers. The addresses, and what they cover, are listed in [Fixed IP addresses](https://pdfmonkey.io/docs/account-and-security/fixed-ip-addresses/).

A word of caution, though: other customers’ documents are rendered from the same addresses, so an IP allowlist **isn’t authentication**. Keep a token or a signed URL on anything sensitive.

## Zapier: no more 30-second race

Zapier gives each action about 30 seconds to finish. With large or complex documents, our **Generate Document** action could run out of time while waiting for the file.

The latest version of our [Zapier app](/integrations/zapier/) works differently: **PDFMonkey lets Zapier know when your document is ready**, and the Zap picks up from there, however long the generation takes. Testing the step in the Zap editor now generates a **real document from your own template** instead of sample data, and invalid JSON gets a **readable error message**. Everything is in the [Zapier guide](https://pdfmonkey.io/docs/integrations/zapier/).

## n8n: version 1.0, verified by n8n

Our [n8n node](/integrations/n8n/) reached **version 1.0**, and **n8n has verified it**. Search for **PDFMonkey** in the nodes panel: an instance owner installs it once, and everyone on the instance can use it, on n8n Cloud and self-hosted alike.

The big change: the **PDFMonkey Trigger registers its own webhook**. Pick a workspace (and optionally a few templates), activate your workflow, done. And when a generation fails, **Generate Document now fails the item** instead of quietly passing along an item with no file.

Upgrading from an older version? Existing triggers need a workspace before their workflow can be activated again, and any webhook you created by hand should be deleted, or your workflow will run twice. The [n8n guide](https://pdfmonkey.io/docs/integrations/n8n/) has the details.

## For developers: a Node.js SDK and an MCP server

The official [`pdfmonkey`](https://www.npmjs.com/package/pdfmonkey) npm package is our **Node.js SDK**, for **JavaScript and TypeScript**. It has **zero runtime dependencies** and runs on Node.js 20+, Bun, Deno, Cloudflare Workers and Vercel Edge. Head over to the [Node.js SDK documentation](https://pdfmonkey.io/docs/integrations/node-sdk/) to get started.

If you work with an AI assistant, the **PDFMonkey MCP server** connects your account to **Claude Code, Claude Desktop, Cursor** or any other MCP client. Ask it to list your templates, generate a document, or check how many documents you have left this month. It runs on your machine and works with your existing templates; it can’t create or edit them. Setup instructions are in the [MCP server documentation](https://pdfmonkey.io/docs/integrations/mcp/).

To let an agent build templates, there’s the **[CLI 2.0](/blog/2026-10-13-pdfmonkey-cli-2.0/)**, which manages your whole account from the terminal.

## That’s it, for now

Thanks for all your questions: they’re how we noticed we’d been a bit too quiet. We’ll do better at telling you about things as they ship.

If something here is new to you and you’d like a hand setting it up, **don’t hesitate to reach out**. We’re here for you!

Happy PDFing! 📝
