I like building things.
Sometimes that means software. Sometimes automation. Sometimes a completely unnecessary idea that somehow turns into a full project.
I come from a WordPress and creative background, and I've gradually moved into software development by building the things I couldn't find elsewhere.
I use AI extensively as a coding assistant — while I focus on the ideas, specifications, design, architecture, testing and figuring out what should actually exist.
Not because I don't care about how things work.
Quite the opposite.
I use AI to help me build. I still want to understand what I'm building.
💡 Ideas I tend to start with a problem, an annoyance, or a thought that begins with "wouldn't it be cool if..."
🎨 Design I care about how things work from the user's point of view, not just whether the code runs.
🤖 AI-assisted development AI lets me turn ideas into working prototypes and software far faster than I could alone.
🧪 Testing Then I try very hard to break what I just built.
🔧 Iteration Build → test → break → understand → fix → repeat.
WhatsApp automation for Hermes Agent
Andoriña started as a personal project to solve a problem I had and gradually grew into something much bigger.
It adds WhatsApp capabilities to Hermes Agent, turning it into a practical, self-hosted automation tool.
It explores:
- WhatsApp messaging & scheduling
- Contact and group management
- Notes and contextual information
- Keyword & semantic alerts
- File and voice-note sending
- Per-contact personalities
- Remote access
- Host-controlled permissions
- Input/output security layers
One of the main ideas behind the project is keeping the host in control.
The LLM can request an action.
The host decides whether that action is actually allowed.
Andoriña is one of the projects I'm working on.
There are more ideas, experiments and projects in various stages — some practical, some experimental, and some that probably should have remained ideas.
I'll add them here as they become real.
