About Chinilla
A simulator for the kind of thinking that happens before the code gets written.
What it is
Chinilla is a browser-based discrete-event simulator for system design. You build an architecture from 7 universal component types, attach 12 programmable behaviors (queues, retries, circuit breakers, rate limits, and so on), and run packets through it. The diagram is alive. You can watch where requests pile up, where they drop, and where the design actually breaks under load.
Why it exists
System design is taught with whiteboards. The whiteboard never pushes back. You sketch a queue, you draw an arrow to a database, you say the word "scalable" out loud, and nobody catches the moment your design quietly stops working. Chinilla replaces the whiteboard with something that runs. Build the design, click play, and the packets either flow or they do not.
What is in the box
- Deterministic discrete-event engine (fixed seed, reproducible)
- 7 universal component types: Person, Step, Storage, Decision, Trigger, Tool, Channel
- 12 behavior modes: passthrough, transform, filter, queue, split, delay, condition, retry, ratelimit, circuitbreaker, batch, replicate
- 16 templates spanning URL shorteners, chat apps, rate limiters, CDNs, and more
- Code-to-diagram for 15+ languages, GitHub repo-to-diagram
- AI design partner powered by xAI Grok (Pro)
- Export to PNG, SVG, animated GIF, Python, Mermaid, JSON
- Mobile PWA viewer with touch-optimized canvas
What it is not
Chinilla is not a load tester for production services. It is not a Kubernetes view, an APM, or a wire-level network simulator. It does not model garbage collection pauses, kernel scheduling, distributed consensus latency, or cache coherence protocols. The numbers you reason with are the numbers you put in. Treat it as a high-fidelity topology and behavior simulator. Use k6 or Gatling to hit real services. Use Datadog to watch them in production. Use Chinilla to think clearly before either of those steps.
Origin
Built by an independent developer. Launched April 2026. Reached top 20 of 569 Product Hunt launches that day, alongside releases from OpenAI, Google, and Anthropic. The project name and aesthetic come from a longer story involving pottery, the moon, dinosaurs, and a particular kind of love. The technical lineage runs through the Collapse Index research on stability scoring for complex systems.
Design principles
- Make the boxes do something. If a diagramming tool only draws shapes, you are still on a whiteboard.
- Determinism over realism. Reproducibility beats high-fidelity randomness for learning and debugging.
- Free tier should be useful. Not a teaser. The whole engine is free.
- Ship the mobile experience. A real PWA, not a "go to desktop" placeholder.
- AI proposes, humans decide. The AI explains and suggests. It does not edit your design without you.