Stigmery: agent-based modelling in the browser, with AI

Stigmery is a free, browser-based tool for building and exploring agent-based models (ABM): models where many simple agents follow simple rules, and complex patterns emerge from their local interactions. Inspired by NetLogo, redesigned around the modern web and an AI that builds the model with you.

What is agent-based modelling?

In agent-based modelling you describe how individual agents behave: wolves, sheep, cars, customers, molecules. Then you watch what happens when many of them interact in a shared environment. It is widely used in biology, ecology, economics, social science, and computer science classrooms to study emergence: how collective behaviour arises from local rules. The name Stigmery is a play on stigmergy, the indirect coordination through traces left in a shared environment (Grassé, 1959).

Build with the AI, or by hand

The AI uses the same tools you do: add an agent type, define a property, write a rule, set up a plot. You can describe what you want in plain English, or edit the model JSON directly. Every change is recorded in an attributed change log, so you always know whether a tweak came from you or from the AI.

Features

• Runs entirely in the browser. Nothing to install.
• Built-in DSL with a TypeScript escape hatch for advanced rules.
• Agents, patches, links, plots, monitors, sliders and switches.
• Predator-prey, flocking, diffusion, segregation: open the classics or build your own.
• Save, fork and share models with a public URL.
• Free to use. Source code on the way.

Looking for a NetLogo alternative?

Stigmery isn't a port. It takes the turtles-and-patches paradigm that made NetLogo so effective in education and rebuilds it for the web, with first-class AI tooling. Existing NetLogo users will recognise the mental model; newcomers can lean on the AI to learn it without fighting syntax.

Dive deeper: read the documentation, browse example models, or skim the FAQ.

Loading the editor… if this message persists, please make sure JavaScript is enabled.