blog

Engineering blog.

Agent architecture, production learnings, security decisions, and cost optimization.

2026-03-20 · architecture, security, open-source

Why We Forked OpenClaw and Built Klawty

How a production deployment with 8 AI agents running a real business led us to build an open-source agent operating system with deny-by-default security.

2026-03-20 · agentic-os, architecture, industry, production

What Is an Agentic OS — And Why We Built One That Actually Ships

Everyone is talking about agentic operating systems. Most are slide decks. Klawty has been running one in production since late 2025 — here's what we learned.

2026-03-20 · production, cost-optimization, case-study

How 8 AI Agents Run a Business for €37/Month

A breakdown of how 8 autonomous AI agents handle 200+ daily tasks for a European SMB — and how 5-tier LLM routing keeps the bill under €40/month.

2026-03-20 · compliance, eu-ai-act, arca

EU AI Act: What It Means for AI Agent Deployments

A practical guide for developers deploying autonomous AI agents in the EU — classification, obligations, fines, and how to prepare before the August 2026 enforcement deadline.

2026-03-20 · security, architecture, nemoclaw

The Security Model: Deny Everything, Allow Explicitly

How Klawty's 3-layer security stack prevents AI agents from accessing the network, filesystem, and shell by default — and why every agent framework should do this.

2026-03-20 · architecture, multi-agent, protocol

Agent Control Protocol: How AI Agents Talk to Each Other

Inside the inter-agent communication system that lets autonomous AI agents delegate tasks, hand off context, and coordinate without a central controller.

2026-03-18 · comparison, architecture, industry

Klawty vs SaaS Automation: Why Zapier and n8n Aren't Agent Operating Systems

Zapier, n8n, and Make.com are workflow automation tools. Klawty is an agent OS. They solve fundamentally different problems — here's when you need which.

2026-03-12 · monitoring, production, observability

Monitoring AI Agents in Production: What to Watch and When to Panic

The health monitor checks every 60 seconds — service heartbeats, DB integrity, failure rates. Here's what to monitor and what the alert thresholds should be.

2026-03-05 · channels, integration, architecture

20+ Channels: Why AI Agents Need to Live Where Your Team Works

Most agent frameworks are API-only. Klawty agents communicate on Discord, Slack, Telegram, WhatsApp, and 16 more channels — because agents that can't talk are toys.

2026-02-26 · autonomy, discovery, architecture

How AI Agents Find Their Own Work

When the task queue is empty, Klawty agents scan their domain for work — emails, GitHub issues, analytics drops. Here's how discovery works without creating spam.

2026-02-19 · governance, proposals, safety

The Proposal System: How AI Agents Ask Permission Before Acting

Walk through Klawty's 6-state proposal lifecycle — from pending to completed with rollback — and the 5-tier autonomy model that decides when agents need approval.

2026-02-12 · security, privacy, pii

PII Detection for AI Agents: Keep Sensitive Data Off the Cloud

How Klawty detects email addresses, phone numbers, credit cards, and IBANs in agent task content — and routes to local models to keep PII off cloud APIs.

2026-02-05 · reliability, dedup, production

The #1 Failure Mode of Autonomous Agents: Spam (And How to Fix It)

Autonomous AI agents create duplicate tasks, repost to channels, and generate identical proposals. Here's the 4-layer dedup system that stops it.

2026-01-29 · architecture, configuration, declarative

AGENT.md: Why Your AI Agent's Config Should Be a Markdown File

The case for declarative agent configuration in Markdown with YAML frontmatter — version-controlled, editable by non-developers, and hot-reloadable.

2026-01-22 · security, policy, yaml

Deny Everything: Writing Security Policies for AI Agents in YAML

A complete walkthrough of klawty-policy.yaml — deny-by-default security for AI agents covering network, filesystem, exec, and PII detection.

2026-01-15 · cost-optimization, llm, routing

5-Tier LLM Routing: Why Your AI Agent Doesn't Need GPT-4 for Every Task

How Klawty routes 80% of agent tasks to cheap models and reserves expensive ones for complex reasoning — cutting AI costs to $1.2/day for 8 agents.

2026-01-08 · reliability, architecture, patterns

Circuit Breakers for AI Agents: Preventing Cascade Failures

How the circuit breaker pattern prevents a failing LLM provider from burning your AI agent budget — with SQLite-backed state and exponential backoff.

2026-01-02 · memory, architecture, qdrant

6-Tier Memory: How AI Agents Remember What Matters

How Klawty's 6-tier memory system prevents token waste and gives AI agents the right context at the right time — from SOUL.md identity to Qdrant vectors.