Hi everyone,
I’m Janis, a product builder and startup co-founder/CTO. Clavus is my independent personal project: an AI assistant for Mac, iPhone, and web.
I know there’s a lot of vibecoded slop around but this is different. I’ve put a lot of time into Clavus, shipped countless updates, and use it myself (and a few friends) for several months already. My focus is making a personal assistant that feels good to use every day, not just in a demo.
It’s open-ended rather than built for one task. A few examples:
Talk through an idea, research a decision, or plan a trip.
Create documents and presentations, then edit them together with your assistant.
Hand off longer research jobs and come back to the results.
Set up recurring tasks, like a weekly check or briefing, or hourly mail cleanup
Pick up the same conversation on your phone, with your documents and preferences still there.
I’m looking for a small group of early testers to try it on their own tasks and tell me what’s useful, what’s awkward, and what’s missing.
I’d offer it for free for now and cover the infracosts myself. There’s no Clavus software or hosting fee during that time, and I’ll provide initial AI credit through OpenRouter, with the allowance confirmed before you join. And you can use your Codex or Claude Code subscription. No automatic renewal or charge. You choose whether to continue.
Afterwards the planned price is $10/month hosted or $5/month for self-hosted updates, with AI usage separate after the initial credit.
It also comes with it’s own iOS Keyboard and dictation features (iOS&Mac) similar to Wisprflow. The whole idea is that when you dictate something it should automatically know about you and what words and names are relevant to you, to detect them properly.
Yeah, it’s hard to explain all the features and I think using it is the best way to explore it. For example, the assistant also helps you discover features that might be helpful to you.
Personally, $120 a year for a vaguely define “assistant” still in beta, in a world where there are already a gazillion AI assistants, when we can use our own AI subscriptions to create whatever assistant we want, is not enticing. There’s no hook that makes the app proposition stand out.
That’s why I provide it for free and even pay the AI costs. I didn’t find a way to convince someone that it works great without them trying it. Definitely something to solve but it’s difficult because everyone can say that, right?
Any idea? Of course I could do a demo video but then again, who has time to watch that.
Happy get advice on what for example you would need to be convinced to give it a try.
I think the challenge might be less about convincing people that the app is well built, and more about giving them one very specific reason to try it.
“Personal AI assistant” is such a broad category now that it’s difficult to imagine how it would fit into my workflow without trying it first.
A few concrete examples like “I used to do X this way, now Clavus does it in Y steps” would probably convince me more than a long feature list.
I’m working on my own software as well, and I’m starting to realize that explaining one small workflow really clearly can be harder than building ten more features.