AI Systems Atlas

Understand AI systems visually.

An interactive atlas of agent architectures, safety patterns, evals, context systems, coding agents, tools, and production operations — drawn as living diagrams instead of prose.

Start with Harnesses See an attack, defended

Seven collections

107 PLATES

Agent Harnesses

The recurring architectures, from a single forward pass to multi-agent teams and agents that run for days.

23 plates · comparison map

Safety & Security

Agents possess authority. The ways untrusted text borrows it — and the boundaries that stop it.

20 plates · attack ⇄ defense toggles

Agent Evals

Outcome checks, trajectory review, judges you can trust — and whether the number is signal.

18 plates · failure taxonomy

Context & Memory

What earns a place in the window, in what order, and how memory brings the right things back.

14 plates · budget builder

Coding Agents

Test loops, verifier ladders, parallel worktrees, and how an edit actually reaches the file.

20 plates · single vs multi

Tools & Protocols

Tool design, MCP, tool search, schemas and parallel calls — and errors a model can recover from.

6 plates · wrapper ⇄ agent-shaped toggles

Production & Operations

Traces, crash-proof workflows, rate limits, streaming, safe rollouts and the bill: agents under real traffic.

6 plates · span waterfall

How to read the atlas

Every diagram uses the same visual grammar. Color is reserved for meaning: neutral by default, blue when something is executing, red where trust ends, green when it’s verified. Press ⌘K to search any concept.

Where this comes from

Engineer's Codex

Real-world software engineering, explained simply. Read by 37,000+ engineers.

Read the newsletter

Zolt Labs

The self-improvement company, and the studio behind this atlas — tools that help people speak better, move better and live better.

Visit Zolt Labs