Introduction

Chinilla is a visual tool for mapping out how a system works. Drop blocks on a canvas, draw lines between them, hit play to watch how things flow, and ask the AI for help when you get stuck.

What can you map out?

Anything with moving parts:

Software (services, queues, databases)
Restaurants (orders, kitchen stations, delivery)
Hospitals (triage, ER flow, who handles what)
Factories (assembly lines, quality checks, shipping)
Logistics (routing, warehouses, last-mile delivery)

You don’t need to memorize patterns or jargon. Chinilla gives you a place to think visually and a way to test your thinking.

How it works

Pick your level of abstraction. Decide what you’re modeling BEFORE you start drawing or asking the AI. The same domain can be modeled at very different levels:

Domain	High level	Low level
A coffee shop	Floor plan: doors, seating area, counter, drive-thru lane	Operations: order placed, drink prep, pickup, payment, restock
A SpaceX rocket	Mission control: launch authorization, telemetry, abort decisions	Engine: turbopump, combustion chamber, nozzle, sensors
A web app	User journey: signup, browse, checkout, support	Backend: API gateway, services, cache, database, queue
A hospital	Building: ER, ICU, OR, pharmacy, admin	Triage flow: intake, ESI scoring, room assign, diagnostics

Both levels are valid; they answer different questions. The high-level “floor plan” tells you what’s connected to what. The low-level “operations” tells you where the bottleneck is and how the system breaks under load. Pick one and tell the AI which (or sketch it yourself with that scope in mind). Mixing levels in one canvas usually produces a confused middle ground that doesn’t simulate honestly at either scale.

Design — drop blocks and draw lines on the canvas (or ask the AI to draft it for you).
Simulate — hit play. The simulator pushes items through and shows where things wait, drop, or break.
Stress test — turn up the volume, fail a block on purpose, slow things down, see what happens.
Analyze — open the Overview panel for health, stability, and bottleneck info.
Revise — fix weak spots by hand or ask the AI, then run again.

What’s in the box

7 building blocks that fit anything: Person, Step, Storage, Decision, Trigger, Tool, Channel.
A simulator that runs the same way every time and shows every item moving through.
Stress tests for peak load, outages, and slow links.
An Overview panel with health, stability, and Monte Carlo runs.
A stress-test panel with one-click scenarios and manual breakage.
An AI chat that can build, explain, and fix designs.
8 ways for blocks to behave (passthrough, filter, split, delay, retry, circuit breaker, batch, replicate) plus queueing on any block.
Per-block infrastructure settings: protocol and scaling (min/max instances, scale trigger).
22 templates: 10 system design interview problems (Pro), 6 lessons, 3 technical examples, 3 process flows (reasoning, support escalation, game economy).
Version snapshots and project history.
Export as PNG, SVG, animated GIF, JSON, System Spec, Python scaffold, Mermaid, or an AI-written PRD.
A guided tour for first-time users.