Architecture designs for the agentic era

A reference library of architecture patterns for autonomous agents, event-driven systems, and AI-native workflows. Read the diagrams, understand the trade-offs, and download what you need.

Explore Patterns

Scroll to explore

The mission

Architecture is no longer just diagrams. It's the operating system of intelligent systems.

In the agentic era, software architecture is fundamentally shifting. Applications are no longer static workflows—they are intelligent, autonomous spaces where agents interact, reason, and act. agentsbutwhy documents the foundational blueprints of this new paradigm.

We believe the best architecture decisions are informed by pattern literacy — understanding not just what to build, but why it works and when it fails.

Why agentsbutwhy

A focused reference for agent-era systems

What each pattern aims to give you — and what it deliberately leaves out.

Layered Clarity

Each pattern breaks a system into clear layers — so you can follow how components connect, not just what they are.

Agent-Native Patterns

Focused on architectures for autonomous agents, event streams, and AI-native workflows — not repackaged web-app diagrams.

One-click Export

Download each diagram as a high-resolution PNG for decks, docs, and design reviews.

Open Architecture

A single, open place for architecture patterns — available to everyone to explore, reuse, and adapt.

Pattern library

Featured Patterns

A few patterns to start with, covering agents, memory, and event flow.

View All

Advanced

MCP Gateway

MCP unlocks tool interoperability, but production systems still need policy enforcement, rate controls, and observability across many heterogeneous MCP servers. The MCP Gateway pattern introduces a governed control plane that brokers MCP traffic, standardizes transport and auth, and gives operators one place to manage trust boundaries.

mcptoolinggovernanceintegrationsecurity

Advanced

Agent-Native Data Infrastructure & Lakebase

Traditional databases were built for humans making careful, deliberate changes. Autonomous agents are the opposite — they branch constantly, test ideas in parallel, and throw most of them away. This pattern shows what a database looks like when it is designed for that reality.

datadatabasepostgresbranchingserverless

Advanced

AIOS — AI Agent Operating System

Running many AI agents on shared infrastructure is like running many programs on the same computer without an operating system — they all compete for resources with no coordination. This pattern applies the same ideas that make operating systems work to the problem of managing agents fairly, safely, and in parallel.

agentskernelschedulingmulti-agentresource-management

Getting started

How It Works

🔍

Explore

Browse a curated library of architecture patterns for the agentic era.

🖱️

Inspect

Pan, zoom, and read each diagram to see how the pieces fit together.

⬇️

Download

Export any diagram as a PNG to share with your team or drop into a doc.

Architecture Philosophy

Designing for the Agentic Future

As systems become more autonomous, infrastructure needs to adapt. We focus on three core principles to build resilient, agent-driven ecosystems.

Decentralized Cognition

Distributing decision-making capabilities across independent, specialized agents rather than relying on a monolithic orchestrator.

Event-Driven Memory

Embracing streams of continuous events as the single source of truth, enabling agents to retain context and seamlessly adapt.

Dynamic Integration

Building systems that can autonomously discover, connect, and negotiate with new tools and external APIs on the fly.

A free reference for builders

Start with a pattern, build from there.

Browse the blueprints shaping the next wave of agent-driven software — free to read, free to download.

Start Exploring