Selected work

AI agents · Product design · Development

Multiagent OpenClaw setup

My own multiagent setup, with a mobile client for voice conversations and a desktop workspace for following the work.

Contribution
System setup, product design & AI-assisted development
Period
2026
Status
Working personal setup · Ongoing development
1:25 · Mobile and desktop demonstration · Voice audio included
agYAP / Louise

A voice for the agents.
A view of the work.

I designed and vibe-coded a custom mobile client for OpenClaw, connected through a cloud gateway. The video demonstrates the capabilities I wanted to add to my setup: real-time voice chat, push notifications and switching between agents. On desktop, I added a deeper view of task dependencies (DAG), files and live agent activity, alongside a pixel-art office adapted from an open-source project.

On mobile
Talk, switch agents, receive updates.
On desktop
Follow dependencies, files and live activity.
01

The situation

I was building my own OpenClaw setup for blog articles, video production, coding and testing. The next challenge was making that system easier to use and follow across devices.

I wanted to talk to agents from my phone, switch between them and receive updates, while keeping a more detailed view of their work on desktop.

02

Why it matters

Delegating work is only part of using an agent system. I also need to understand what is running, inspect the outputs and decide what to do next.

A mobile interface makes the system available away from the desk. A desktop workspace provides the detail needed to review dependencies, files and ongoing activity.

03

My contribution

I configured the multiagent workflows and designed and built the interfaces around them using AI-assisted development. The mobile client, agYAP / Louise, supports voice and text conversations in English and Czech.

I connected the client through a cloud gateway and added push notifications and agent switching. On desktop, I added views for task dependencies, files and live agent streams, with a pixel-art office adapted from an open-source project.

04

Key decisions

Give mobile its own interaction model

I designed a dedicated mobile client around real-time voice and text, with agent switching and push notifications.

The result is a working phone interface for conversations and updates, demonstrated in the recording above.

Keep deeper inspection on desktop

I added separate DAG, Files and Live views for following task dependencies, reviewing outputs and inspecting agent activity.

These views bring the underlying work into the interface alongside the conversation, making individual tasks and outputs available for review.

Build on existing foundations

I used OpenClaw for the agent system and adapted an open-source pixel-art office for its visual workspace.

The custom work sits around those foundations: configured workflows, the mobile client and the monitoring interface.

05

The outcome

Working personal setup · Ongoing development

The result is a working personal agent setup with configured workflows for content, video production, coding and testing, plus a custom mobile client and desktop monitoring.

Čistý kout was the first project I built using this agent system. The setup and its interfaces continue to evolve as I use them in my own work.

OpenClawAI agentsMobile UXReal-time voiceWorkflow orchestrationAI-assisted development
Next projectHiFiveUI
Pavol JuríkProduct / UX / AI

Let’s talk

Good things start
with a conversation.

Have a role, a project, or an idea in mind?
Tell me a little about it.

Just a conversation.
Your details stay between us.

Prefer your own email?contact@pavoljurik.com