Welcome, ICC attendees — you found the room where it happens first!
This is the Ignition2027Early Access forum, and it's exactly what it sounds like: the front-row seat for everything coming in the Ignition 2027 release. This is where the beta drops, and if you subscribe here, you'll be amongst the first to know.
A lot is coming in 2027 — and we mean a lot. We're not ready to pull back the curtain on the specifics just yet (some of the best stuff you'll hear about right here at ICC), but we can tell you this much:
There's a beta on the horizon packed with fresh new features and enhancements.
Some of the things you've been asking us for? Let's just say you'll want to be watching this space.
A few surprises are in the mix that we think will genuinely change how you build.
Subscribe to this category now — hit the bell / notification button at the top of the page, and the moment the beta is ready, it lands straight in your inbox.
We can't wait to show you what we've been building. See you on the inside!
Anyone else underwhelmed by the Catalyst announcement?
I can already do this today, with my Claude Code subscription and the Artek MCP module. Don't even have to open Designer anymore, only to verify the changes.
So who's actually going to pay per token for development using a LLM API? Or have a machine running a local LLM big enough to make it worth it?
Assuming this module will not be free, was there any announcement on pricing?
I have been working with IA since 2009 - Gold Certified for a long time - and just don't see me using this. I was not at ICC this year and they didn't have remote tickets so don't know the full details yet.
The Catalyst "core" stuff, everything meant to aid development in Ignition, is free.
The custom user agent stuff, that let's you develop custom agents that run in your deployed projects, will cost something - not announced or known to me.
I'm actually excited about it. I was at ICC and having the AI baked directly into designer and the gateway web interface is a game changer. Being a first party solution, you can just prompt it for various things like creating views, SQL queries, tags, UDTs, scripts directly in the windows you already use in the designer with auto-complete functionality so you can tell it which component to get the data from, or where to write it to. On the gateway the diagnostics analysis will be great for those more difficult errors so instead of searching through logs, thread dumps, etc, the catalyst module will do all that for you and find the issues with suggestions on how to fix it (if it knows).
Understood - My point is that I can do all of that today using a $20 Pro subscription. I have had CC get live thread dumps and view live logs and give analysis that has helped me find stuff I probably would not have found. I have it creating views, UDTS, Instances and scripts daily with an Ignition Skill that has been getting really good the more I train it.
Maybe not everyone is as good as you, and there could be a large set of people that could benefit from pre-defined tools and skills that help them in designer without having to become AI experts first?
Not every SCADA feature has to be adopted equally be every engineer, nor will every engineer find certain features equally as useful as other engineers.
Dropping an MCP server into the Gateway scope, or directly editing resource files, is a hack to work around not owning or being able to modify the platform to support AI in a first class way. The scope is going to be limited compared to what we can do in the platform/designer, and it's simply not (as) safe.
The Designer is the obvious and safer place for this to be happening, where you can preview changes before they're irreversibly (without git revert or something) applied to your resource files, and we can provide all the tools needed for the agents to do their work.
We're eventually going to add an MCP server to the Designer itself, with the intent of letting people hook their own agents/subscriptions up to that, but I'm not really sure on the timeline or if it will be as effective as the built-in AI functionality.
It's going away because its functionality was folded into the Catalyst module.
edit: I assume most people didn't ever bother to use or set it up, so this may seem contradictory to what I said in a previous post. The preview MCP module was basically just WebDev for MCP. It did nothing by default. It was just a blank canvas for you to define tools. It was not a set of Ignition-related MCP tools that we authored or provided.
I'm not sure how the Med device industry will adopt this with heavy regulations to validate and lockdown code. I can't see an adoption where code is changing by agents.
I agree the Designer is the right place, especially for previewing and rolling back changes.
I don't think my current workflow is a "hack" though. Most of it goes through an MCP module built on the SDK. The file-level edits that do happen are in a git-tracked project on a local dev gateway, and I review / test every change, line by line, in the Designer before ever exporting to production.
Excited for a Designer MCP module. Hoping it comes out soon.
Looking forward to this. Catalyst is amazing. Native excel tag import/export looks so good. Can’t wait to get my hands on the beta. Also agree with richardNZ15's point. A lot of engineers on our teams won't set up their own MCP servers and custom skills, so having this built into the Designer, with approval and rollback, is what will actually get AI adopted on real projects.
Is there anywhere I can watch the keynote talking about the 2027 release features? I'm seeing a lot of comments on my socials, but I'd like to hear what was said specifically at ICC.
Subscribe to this category now — hit the bell / notification button at the top of the page, and the moment the beta is ready, it lands straight in your inbox.