The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
I built ClaudeMate because I kept missing the moment a session needed me.
Last year I shipped eight production platforms solo. That works because I run several Claude Code sessions at once. It stops working the moment you have to babysit them. One is mid refactor. One is quietly waiting on a decision you made an hour ago somewhere else. One went off the rails forty minutes back and you find out at review. So you tab between terminals all day, and the tabbing becomes the job.
ClaudeMate is a small piece of hardware sitting on my desk. A USB device with Arduino firmware and a Python daemon that hooks into Claude Code and wraps the PTY. It watches the live sessions and physically signals when one needs a human. That is the whole product. Open source, MIT licensed.
Now the part I actually want to argue about.
Every tool in this space is optimising tokens. Smaller context, better compaction, cheaper model per call. Good. Necessary. Also only half the bill.
The other half is the person spending them. Context switching is not free. Neither is holding four half finished threads in working memory, or making the twentieth micro decision of the day when your judgement is already thin. Token cost you read off a dashboard. Operator cost you read off the quality of your last commit of the day.
Which is exactly why the answer was hardware and not another dashboard. A dashboard is one more thing you have to check. A light on your desk is a thing that checks you. It takes the cost off your attention instead of adding to it.
An agent setup that saves you thirty percent on tokens and costs you your attention has not saved anything. It moved the invoice somewhere you cannot see it.
Who it is for. Solo engineers and small teams who run more than one Claude Code session at a time and treat it as a production tool, not a toy.
github.com/vlgutv22/claude_mate
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Trending
retrofuturism
Share your favorite designs from the past exploring the distant future. Or your own take on the aesthetic 🚀
halftone
Dots. Texture. Grain. Let's see your designs using the halftone technique.
Midjourney
Creatives are showing off what they're making with Midjourney.
colorgrade
Video editors sound off: How do you enhance the mood of your video in color grading?
Related posts
Okay, this is wild. I built a tool today in Claude that's going to save me days of work, and it took about an hour of playing around.
I'm building out SOTO, a concept wearable brand, and I needed a fast way to test how our key visual moves across every place it'll show up. So I asked Claude to build me a tiny motion design site. Pick a frame size. Drop in the visual. Play with the animation settings. Export a clean loop. That's it.
No more re exporting from After Effects for every single format. No more guessing how something feels at 1:1 versus 9:16 before it even exists. I just get to see it, right away.
Honestly, this is the part of AI that gets me excited. Not replacing the creative call. Just clearing out the annoying steps between having an idea and actually watching it move. Without Claude I would have sat on this motion design aspect for some time and maybe not even get to it with other projects knocking not the door.
Curious if anyone has exciting case studies they could share using this same approach?
That is really incredible! Are we gonna be able to reach SOTO ?
Everyone loves to say AI makes creative work feel impersonal but I found the uses that actually save time without replacing my creative expertise.
Here's what's been working for me:
✍️ Copy and emails: I write the base, AI helps me polish and tighten it
🎨 Concept and moodboard copy: same approach, I write the core idea, AI helps shape the language for presentations
📦 Mockups: AI speeds up creation so I don't spend hours looking for the right .psd online
The key has been using it to speed up the boring parts so I can focus on the creative bits.
What's an AI use in your workflow that actually made things better, not just faster?
Starting with your own draft is close to how I use it on builds. I give AI the page and a short checklist, then use what it catches to tighten the work before I ship.
Recently I tried something new: I became a founder of my own pet project.
Over my 6 years in design, I've worked with all kinds of companies: from big tech to startups, across very different business types. Some clients were in love with their product and wanted to build only their own ideas, so I leaned hard on soft skills to hold space for both the business and the users. Others were open to suggestions and design-led decisions.
From a designer's side, it's always a challenge and a balancing act. But when I started my own pet project (a product for expat women), something shifted. This time I was the one fully in love with the idea, and I wanted to learn vibe-coding and Cursor hands-on.
With Claude, I quickly shaped the idea, positioning, competitor analysis, and other pieces of product research. I even validated my hypotheses through surveys on Reddit, Facebook, and with my friends — and they held up. Happy and motivated, I jumped into building the landing page and core flows in Cursor…
…until I caught myself. I was about to sink a lot of time into building, and this time, there was no team whose hours I needed to justify but my own.
So I stopped and asked: what's the cheaper, faster way to test this? I had two options:
Build a Telegram bot — faster than full development.
Test manually and gather feedback from participants directly.
Here's the thing: cheap, early hypothesis-testing wasn't new to me. I knew it well, but always from the designer's seat, as one good practice among many. What changed with my own project is that the trade-off finally became mine. It wasn't a team's roadmap or a client's budget on the line, it was my own time, my own money, my own idea I was tempted to over-build.
And that's exactly what made it click, the way the Lean literature I'd been recommended frames it (The Lean Startup by Eric Ries, Lean UX by Jeff Gothelf): the cheapest way to test a hypothesis before you build isn't just a nice principle. As a founder, it's the best decision.
Being on the founder's side, even briefly, shifted what I want to get better at. Understanding founders: their state of mind, what drives them, why they're so attached to building things a certain way. And from there, being the person who can offer the simpler, cheaper path before anyone sinks weeks into the wrong thing. Now that I've felt that pull to over-build myself, I understand it from the inside, and I think that makes me a better partner to the founders I design with.
If you've made the leap from designer to founder (even for a side project), I'd love to hear what shifted for you.