“Even when I’m eatin’, I be on alert.
I ain’t with the frontin’, I just make it work.”

— Larry June, The Good Kind

It’s been exactly one year since posting an update for Hypermatic’s 5th birthday (From Four to Five), so here’s another annual recap to celebrate Hypermatic turning 6 years old today!

Some numbers to celebrate from 2025

• 0 investors and $0 funding (still happily bootstrapped/default alive)
• 17 new YouTube tutorials recorded
• 432 new product updates and improvements shipped
• 1,807,900+ users (up from 1,450,000 a year ago)

Some new feature highlights shipped in the last 12 months (since January 2025)

Besides the usual bug fixes, customer requests, and other smaller enhancements, these were some of the other new features that made their way into the @hypermatic feed on X (which is mainly used to keep a record of these changes) over the last 12 months:

• Added Multi-Platform HTML Exports to Bannerify
• Added Email Template Categories to Emailify
• Added Advanced Merge Tags to CopyDoc
• Added JPG/PNG Slide Exports to Pitchdeck
• Added Outlook Email Signature exports to Emailify
• Added Figma Buzz support to CopyDoc
• Added Figma Buzz support to TinyImage
• Added Placeholder Text Syncing to CopyDoc
• Added Animated PNG (APNG) Exports to TinyImage
• Added Saved Custom Code Presets to Bannerify
• Added Animated PNG (APNG) Exports to Bannerify
• Added Custom Filenaming Updates to TinyImage
• Added HTML Text Optimisations to Emailify
• Added PowerPoint Video Embeds to Pitchdeck
• Added Self-Hosted Web Presentations to Pitchdeck
• Added Component Variant Syncing to CopyDoc
• Added Login Page Customizations to Pitchdeck
• Added Divider Line Padding to Emailify
• Added Text Color Backgrounds to Emailify
• Added Text Padding to Emailify
• Added Export Presets to Emailify
• Added Video → GIF imports to Convertify
• Added Custom Image Hosting to Emailify
• Added new Figma UI3 style updates to all Figma plugins
• Added a Bento export option to Emailify
• Added Hybrid Previews to Emailify
• Added Figma Slides support to TinyImage
• Added Figma Slides support to CopyDoc
• Added a Yotpo export option to Emailify
• Added Figma Slides support to Pitchdeck
• Added Feedback Filtering to Commentful
• Added Split File Exports to CopyDoc
• Added BidTheatre exports to Bannerify
• Added Kanban Boards to Commentful
• Added Font Overrides for Google Slides to Pitchdeck
• Added Layer Hiding Spreadsheet Sync to CopyDoc

Skipping the 2025 mid-year update post

For the last six years, I’ve always done a mid-year “Independence Day” update on July 4th; and while I started drafting that for 2025 at the time, there was a bunch of other stuff going on that made it feel weird to post an update while I was so unsure about a bunch of things that were up in the air at the time. I will try to cover the entire year in the post below instead, with some personal and business updates, as both sides have influenced the direction of the company.

Health scare (with another family member) in January 2025

At the very start of the year, without going into personal details, there was an unexpected and sudden health scare with a close immediate family member that very narrowly escaped death.

I think that this kind of set the tone for the rest of year, where I started thinking more about how I was spending my own time, and reflecting on some changes I wanted to make for myself personally, and also within the company going forward. You will see some of these things in the following sections below.

Removing our Enterprise offering

Earlier this year, I made the decision to completely shut down our Enterprise offering. This was a very intentional decision to move away from the Enterprise space as a company-wide direction.

Of course, this decision came at a pretty large financial cost, but the trade-offs have been worth it for how I want to run the company, rather than just following the well-known path that most software companies end up following, where they “double down on Enterprise sales”; and I don’t want to do that, because it’s not the right fit for Hypermatic.

Since then, this has meant turning down customers with requests for these services (at a higher price point), but staying on that path has helped us stay nimble, focused, and able to continue to offer lower pricing to all of our other customers on the self-serve Pro license.

The most contrarian thing of all is not to oppose the crowd but to think for yourself.

— Peter Thiel, Zero to One

As some more context, for anyone who has not been through the Enterprise sales cycle, there’s an accurate article called Bending over: How to sell to large companies that’s a good introduction.

Basically, this is a very complex ongoing process that usually takes a minimum of 6 months per new customer, but can take as long as 18+ months (in more extreme cases) from the date of the first email enquiry to signing the final agreement.

We started supporting this earlier on in the company’s history to try and help accommodate very specific requests from certain customers, but over time, this took up much more time, energy and bandwidth, and was (by far) the most de-motivating and unpleasant part of running the business.

The final straw that led to completely shutting down our Enterprise offering came after spending almost 12 months after doing all of the pre-work and onboarding requirements for a future Enterprise customer.

The process started with all of the usual things — filling out a bunch of forms to be onboarded into a vendor system/platform, sending/receiving dozens of documents, passing security questionnaires with hundreds of questions, completing various surveys about environmental impact, agreeing to diversity and inclusion quotas, and signing documents verifying that we are indeed against child slavery.

However, one of the later steps involved taking out a very specific insurance policy just for this one customer, only to be completely “ghosted” soon after, where we never heard back from them again.

While this particular example was an extreme case, in terms of its (non) outcome, the process itself of going through all those steps for multiple companies at the same time is not how we want to be using our energy.

It has been said that “You can just do things”, and as the founder and CEO of Hypermatic, this is one thing I’ve decided to just do (or not do, in this case), and I can confirm that it has been one of the best big directional decisions that I have made so far.

Not selling the company (and remaining 100% independent)

As I mentioned, and as you can probably tell from my extended “rant” above, the Enterprise side of things was really wearing me down earlier this year, to the point where, for the first time ever, I actually entertained the idea of selling Hypermatic as an alternative option during the middle of 2025.

We got all the way up to the stage of receiving a formal offer from one acquirer, but I felt in my gut that it wasn’t going to be the right fit for the future home of Hypermatic, and so I graciously turned down the offer; once again, at a large financial opportunity cost (if you’re picking up on a recurring pattern).

After deciding against selling, as it always had been prior to then, Hypermatic remains 100% independent, self-funded (bootstrapped), and has still never taken any outside investment to this day.

In the end, while the Enterprise side had worn me down and dampened my day to day enjoyment of running the business, I decided that I still love working on the company, our products, and our customers; so I realised that the goal should be trying to make the business the right “founder-fit” for me personally, and in turn, helping to continue deliver the most value to our customers as possible.

Making our company-wide policies public

After deciding that I need to make some changes to ensure that I’m building the company that I personally want to work on every day, and cutting the Enterprise tier from our offerings, I also decided to draw a line in the sand for a bunch of other things, too.

Rather than having to decide on a case-by-case basis, or continually reply to the same enquiries (only to say that we can’t help them with their request), I decided to actually write these down and formalise them as company-wide policies that are easily visible on our contact page, before we waste anyone’s time with reaching out about them:

• We don’t have meetings
• We don’t jump on calls.
• We don’t offer discounts.
• We don’t fill out forms or surveys.
• We don’t go through vendor onboarding.
• We don’t accept bank transfers.
• We don’t white-label our products.
• We don’t do partnerships.
• We don’t do affiliate programs.

Similar to cutting our Enterprise tier, I have no doubt that these policies will cost us opportunities, but the trade-off is worth it. As I wrote on our contact page below:

“We’re an independent calm company that likes to keep things really simple — we build excellent products and provide amazing customer service at a fair price — that’s all we do. We only work with customers who share our values and who also like to keep things simple. If that’s not you, we’re not going to be a good fit for each other, no hard feelings!”

Just because some customers would prefer us to do some of these things, doesn’t mean we have to conform our own company culture to facilitate those things to fit their own company culture policies.

There’s no law that says you need to agree to a 1-hour meeting invite with half a dozen people for something that could easily be answered in an email, and I think everyone on both sides would be thrilled to have that hour of their finite life back, in any case.

Not to get too deep before cocktail hour, but need I remind you of the finite nature of life?

— Roger Sterling, Mad Men

The reality is that it’s incredibly easy to just say “yes” to every request, to the point where that ends up becoming the only things you spend your time on; if we agreed to every type request above, nothing would get done, because we’d be on Zoom calls or trapped in a SAP Ariba dashboard hellscape filling out 4,000 forms for months on end, and have no time or motivation to actually add value to our customers by building things they need, or providing them with excellent customer support.

Claude Code joins the team

Another really big change that also influenced the future of the business was the release of Claude Code earlier this year. The short summary is that Claude Code is already so good (and getting better every month) that I will not be looking to hire any developers in the future to help with writing code. Despite what you may have heard, it is not just a “junior developer” or “good for small tasks”, it is much closer to the experience of collaborating with some of the best human developers that I have ever worked with.

After giving a bunch of different AI enhanced coding tools a try over the last few years (Github Co-Pilot, Cursor, etc), it wasn’t until I gave a proper try earlier this year that it finally clicked, where I had the same feeling that I had seeing Figma for the very first time back in 2018. It is the feeling where you could instantly tell that this was clearly going to be the future of a workflow; in that case, it was the future for design, and in this case, it’s the future of development.

Like any software company, there is always a list of features, improvements or customer requests that have taken a back seat; adopting Claude Code was the perfect tool to help with all of this, as it’s very good as refactoring code, and working on very tightly scoped features or improvements. As of now (December 2025), with the help of Claude Code, I can say that the backlog that existed at the start of the year has almost been entirely completed.

There is lots of talk about an “AI bubble”, which I won’t speculate on here, but all I can say that is that Claude Code (and other coding agents) are a complete game-changer, and will be the default way that 99% of all code is written in the future.

It’s interesting to still see so much skepticism, negativity, or outright hate in terms of these AI tools. It seems like there are some people who are totally blind to how good these tools are already (and how they will continue to improve in the future), because they have already positioned themselves as being opposed to them for other reasons. It makes me feel like I am living in a parallel reality when I read articles like this from September 2025, that open with the paragraph below:

“I think it’s long past time I start discussing “artificial intelligence” (“AI”) as a failed technology. Specifically, that large language models (LLMs) have repeatedly and consistently failed to demonstrate value to anyone other than their investors and shareholders. The technology is a failure, and I’d like to invite you to join me in treating it as such.”

At the same time, I have Claude Code completing features or code improvements perfectly and autonomously, where I can review every line of code it added or updated, just like how you would review the code of another developer on your team before accepting their pull request and deploying the code (after testing it).

In a world where you can already use natural human language to tell a tool like Claude Code on what you’d like it to do (just like you would describe to another human developer on your team), and have Claude Code complete that task as well as you (or your co-worker) could; it seems wild to say that this fails to demonstrate value to anyone besides investors or shareholder of AI companies, rather than the users of these AI tools, who can benefit from using them for free, or at very low cost.

Even more confusing is the final line of the article above:

“‘Artificial intelligence’ is a failure. Let’s you and I make sure it stays that way.”

Again, this seems like a “tell” that this way of thinking it’s less about the technology itself, and more about a certain worldview or identity that isn’t compatible with AI succeeding; to the point where actively ensuring the technology isn’t allowed to succeed is the goal, rather than figuring out how to make it better.

If AI is truly a “failure” and doesn’t work, at least in terms of LLMs for coding, why has Claude Code reached $1 billion in run-rate revenue in only 6 months? This is the fastest product in history to go from $0 to $1 billion, so it seems more likely that it does actually “work” (really well), and everyone who is voluntarily paying for it are getting way more value from it than the $20/month - $200/month it costs them.

I think betting against AI progressing and improving is not a great strategy. This is where much of the investment capital and talent is going; of course, just like the early internet dot-com bubble, there will be too much hype, maybe too much pre-mature build out, many companies who will not make it, and probably a “correction” in some of the AI related stocks. All that said, it seems obvious that AI will be one of the driving forces of progress in the coming decades, and I think being optimistic about all the positive things it will be able to do is much healthier than being too black-pilled and willfully ignorant about what’s already happening right now, and it’s only going to keep getting better.

Migrating the marketing site to Astro

After using the Gridsome static site generator since 2019, which had its development stopped a few years ago, I decided to migrate this website to the Astro static site framework instead.

Because Gridsome was Vue.js based, and Astro uses its own .astro files (although you can use Vue if you prefer), I used Claude Code to help port the previous Vue components over to work natively in Astro and remove the Vue.js dependency entirely. After migrating all of the content, pages, styles, and interactivity over, the result is a new static HTML website that is even faster than Gridsome was.

Migrating to Astro was also a the perfect opportunity to ensure all of the content was up to date, and do some re-design work in terms of typography, colors, image assets and copywriting.

Spending time Taiwan and South Korea

After working on Hypermatic every day since December 2019 without a full day off, and not traveling during that time (partly due to Melbourne being in lockdown for over 2 years), I decided that after five years of building the business full-time in Australia, it was time to take advantage of remote work and step out of my comfort zone while I still have the time, health, and flexibility to do so.

Life moves pretty fast. If you don’t stop and look around once in a while, you could miss it.

— Ferris Bueller, Ferris Bueller’s Day Off

To force myself out of my comfort zone, I decided to visit Taiwan in March, and then South Korea in October. I had a really great time in both places, and I will certainly go back to visit both again in the future, as there’s so much to see.

Taiwan (March 2025)

In March, I spent 3 weeks in Taiwan, where most cafes don’t open until after 10:00am, so I would usually either work from my hotel, and then try a different specialty cafe to work from later in the morning. I still worked every day, but was able to see a lot of Taipei, Tainan and Kaohsiung.

Taiwan seems very underrated as a tourist destination. I barely hear anyone talking about going there, and it wasn’t uncommon to go many days without seeing another western foreigners there, especially after leaving Taipei.

Coming from Melbourne, which is well known for having great cafes and coffee, I was surprised to see how thriving the coffee scene was in Taiwan as well. I came away realising that Melbourne does not have the best coffee in the world (although it’s usually quite good), as I previously assumed we did.

There are many specialty coffee shops, and the owners clearly take an extreme level of care in making it for you. I think the best coffee I’ve ever had was at the Simple Kaffa flagship store in Taipei; I went there 3 different times during my stay, where they always let me work for most of the morning while having their regular coffee, along with trying some other options like infusing whiskey or Taiwanese tea, which were also excellent.

There is very little English spoken or printed on menus, but I was able to get around easily enough and try a bunch of popular food there: Beef Noodle Soup, Xiao Long Bao (soup dumplings), Hu Zhao Bing (pork pepper bun), Lu Rou Fan (braised pork rice), Gua Bao (pork belly bun), Cong Zhua Bing (scallion pancake), Danbing (egg pancake roll), Taiwanese shaved ice, and of course, Bubble Tea.

South Korea (October 2025)

In October, I spent 3 weeks in South Korea, where I was still working 3-4+ hours per day, also still working every day, but I still had more than enough time to explore Seoul and Busan during the day and evenings.

Compared with Melbourne, which has felt a bit like its on life support since 2020, being in Seoul (and around Hong-dae in particular) felt like what a city should feel like. There was a huge amount of energy, everyone was out and about, being social with others and seemed to generally be having a good time.

Being in South Korea, it was very striking how different the culture was compared to Australia. It seems somewhat nationalist and inwardly focused. It feels like there’s very little outside influence or care about what other places are doing. Many of the brands you use (cars, electronics, etc) are all Korean, the advertising you see is all Korean, the music you hear everywhere is all Korean. Similar to Japan, everything is also very functional, and more modern. People there were super friendly and helpful.

Like in Taiwan, I was also very surprised at the scale and quality of the coffee scene in Korea. There are many coffee shops, apparently over 100,000 across South Korea, and over 20,000 in Seoul alone. Similar to my experience in Taiwan, it’s clear that South Korea has also definitely surpassed Melbourne in taking coffee very seriously, at specialty shops and roasters, at least.

Overall, I think South Korea was my favourite trip out of the few places I have visited outside of Australia so far.

Looking ahead to 2026 for Hypermatic

Looking back on my 2024 end of year update, I used ChatGPT to summarize the things I wanted to deliver in 2025 into the bullets below, which I’ve marked as a success or miss.

• ✅ Prioritise stability, quality, and improvement over feature creep and UI bloat
• ✅ Focus on behind-the-scenes upgrades, performance, and better defaults with minimal UI disruption
• ✅ Ship integrations and enhancements that fit naturally into existing Figma workflows
• ✅ Publish more free design resources that work out of the box with the plugins
• ✅ Create more free video tutorials to help users unlock deeper value
• ✅ Migrate the marketing site to a modern framework and audit/simplify content
• ✅ Improve and maintain backend systems and internal tools
• ❌ Explore a new tool for building sites and landing pages that exports production-ready, user-owned code
• ❌ Use LinkedIn more deliberately to increase awareness within your existing network
• ✅ Maintain a long-term, sustainable focus on product quality, trust, and real user problems

The theme for this year was about taking a pause on making radical changes to the Figma plugins, and focusing on improving the performance and quality of them, without causing too many UI changes.

Instead of continuing to bloat them with new features for the sake of it, I wanted to make the existing features more valuable by adding more platform integrations and quality of life improvements.

In terms of building a new plugin that would help users create websites and landing pages, after having this on my “wishlist” for many years, after seeing how Figma executed its “Figma Sites” product and its new “Figma Make” product, I just can’t see a huge opportunity for being able to deliver additional value to what those two already offer natively.

I do have a few ideas for brand-new Figma plugins that I am keen to explore in 2026, and I think going back to our roots of shipping something early and figuring out what is going to deliver the most value (then doubling down on those) is still the right approach.

Besides that, there are still a bunch of other improvements and features for many of the existing Figma plugins that would be great to add over the next 12 months as well, and I’m always surprised to learn more new use cases from our customers (which lead to new improvements and features) every week, so I’m sure there are still many things to come in 2026 that I haven’t even thought of yet.

Next year, Hypermatic will also be going to the Figma conference (Config 2026) in San Francisco,for the first time. It will be cool to meet some customers in person, learn about the new Figma product updates in real-time, and also just generally get a sense of how people are feeling about design and development in mid-2026, which can be hard to tell just by trying to track it online (on X or Reddit etc) or by chatting with customers via email.

All in all, after working on Hypermatic every single day for the last 6 years, since December 21st 2019, I am incredibly thankful that I still get to do this full-time. My simple (and somewhat naive) dream when I first started was to be able to build design automation tools as a small self-funded startup, and be profitable enough to pay for my living expenses and continue working on helping to solve these problems for customers indefinitely. My assumption was that there were other people out there who were facing similar workflow headaches to the ones that I had when I worked as a designer/developer for 10+ years, and that mostly turned out to be true.

With that said, thank you to everyone who has tried out any of the Figma plugins from Hypermatic over the years, and I hope it has helped ease or remove some of the pain that you would have otherwise felt in your own design workflows.

I will be sure to post a mid-year update next year, which will be published shortly after getting back to Australia from the Figma conference in San Francisco. Wishing everyone who made it reading this far down the post (all 3 of you) a great start to your 2026 in the meantime!