LivreMente

Skip to content
Case Study — AI Agent System

My team of AI agents

4 autonomous agents that publish, research, prospect clients and improve themselves — 24 hours a day, 7 days a week.

I want a system like this

4

autonomous agents

56

active cron jobs

9

platforms

24/7

always running

The System

Ruben talks directly to each agent. Nebur coordinates, but also executes — and agents escalate to each other when needed.

👤 Ruben
🧠NeburLeader
🌱SeedContent
🔍ScoutNews
🎯HunterLead Gen

Ruben also talks directly to Seed, Scout and Hunter

What they do

Each agent has its own specialty and personality. They share a common mindset. They work independently — and collaborate when needed or when Ruben asks.

🧠
NeburOrchestrator

The boss. Coordinates the entire team, runs strategic analyses, writes code, builds websites and solves complex problems. Runs weekly retrospectives and continuously improves the system.

CoordinationStrategic analysisCode & websites
🌱
SeedContent & Publishing

Publishes across 9 platforms, 4 times a day. Creates quotes, clips, BIP posts and news. Monitors performance, runs A/B thumbnail testing and auto-approves content when there's no rejection within 2 hours.

9 platforms20 tareas programadasSmart auto-approval
🔍
ScoutResearch & Intelligence

Researches news across LivreMente's 6 pillars. Cross-references RSS, Reddit, Brave and X. Learns from every validation and rejection. Self-improves weekly.

6 pillarsFeedback loopSelf-improvement
🎯
HunterLead Generation

Prospects clients who need AI automation or an online presence. Searches the web directly (Google Maps, local directories, social networks) and platforms like Upwork, LinkedIn, Reddit and IndieHackers. Prepares personalised proposals. Never contacts without approval.

Active lead genPipeline trackingDraft proposals

How it was built

4 agents. A system in constant evolution.

Phase 1

Nebur as personal assistant

First agent — various tasks, code, organisation, strategy.

Phase 2

Seed is born

Need to scale content. Seed specialises in autonomous publishing across 9 platforms.

Phase 3

Scout as journalist

Continuous news monitoring across LivreMente's 6 pillars. Cross-references sources, validates, submits.

Phase 4

Hunter for lead generation

Active prospection on the web (Google Maps, local directories) and platforms (Upwork, LinkedIn, Reddit, IndieHackers) — clients who need a website or AI automation.

Today

56 tareas, 22 habilidades, 58 scripts

A system with feedback loops, weekly self-improvement, approval by silence and shared memory between agents. Top 5% of autonomous agent setups globally.

Under the hood

Not a chatbot. A production infrastructure.

Multi-agent with hierarchy

4 specialised agents with clear roles. Nebur coordinates, the others execute. They escalate to each other when needed.

Persistent memory

Each agent has long-term memory, diaries, feedback loops and an anti-compaction system that preserves critical context.

Weekly self-improvement

All agents analyse their own mistakes and propose improvements to SOUL.md and workflows. The system gets better on its own.

Approval by silence

Content auto-publishes if not rejected within 2 hours. The human only intervenes when they want to. The system never stops.

Browser automation

Agents control a real browser to publish, research and interact on platforms that don't have an API.

Production-grade security

Isolated Docker containers, read-only, hardened SSH, encrypted secrets. Never acts without approval on critical actions.

Real results

Not theory. This runs every day on my server.

~100

agent sessions per day

Each cron job is an autonomous session

22

specialised skills

Publishing, SEO, thumbnails, research...

58

automation scripts

Navegador, analytics, tokens, cookies...

4x

content published per day

Quotes, news, BIP, clips

6

research pillars

Scout researches 2x/day on each pillar

0

intervention needed

Approval by silence within 2 hours

For your business

Want a system like this? I can build one tailored to you.

Content Automation

Full pipeline: creation, smart approval and publishing across 9+ platforms on autopilot. The same system I use daily.

€990 + €300/month

Popular

Multi-Agent System

Up to 3 specialised agents for your business: content, lead generation, research. With feedback loops and built-in self-improvement.

€2,900 + €700/month

Full Autonomous System

4+ agents with hierarchy, persistent memory, weekly self-improvement and dedicated infrastructure. The same system I use.

€5,900 + €900/month

I want a system like this

See all plans and pricing

I reply within 24 hours.

hello@rubenmoderno.xyz

Frequently asked questions

How long does it take to set up a system like this?

Content automation: 1-2 weeks. Multi-agent system: 2-4 weeks. Full autonomous system: 4-6 weeks. Depends on integration complexity.

What if the agents make mistakes?

You define the level of autonomy. Each action can be automatic or require your approval — publishing, contacting people, spending money. The agents have circuit breakers, iteration limits and full logs. You always have the final say.

Which platforms do they support?

Instagram, Facebook, X/Twitter, LinkedIn, TikTok, Threads, Bluesky, YouTube, Substack, Medium — and any platform with an API or accessible via browser.

Do I need technical knowledge?

Zero. I set up, configure and maintain everything. You interact with the agents via Telegram, WhatsApp or your preferred platform.

How much does infrastructure cost?

Server + AI tokens run around €50-150/month depending on volume. This cost is separate from the maintenance service.

Can I start small and scale later?

Exactly. Most clients start with Starter (1 agent) and scale as they see results. No long contracts.

Ready to automate?

Send me an email with what you need. I'll reply within 24 hours with a concrete proposal.

Talk to me

hello@rubenmoderno.xyz