dcs-simulation-engine 0.4.0

A textual environment for simulating diverse cognitive systems.

Diverse Cognitive Systems (DCS) Simulation Engine

A playground for interacting with diverse cognitive systems.

What is this?

A gameplay framework for interacting with diverse cognitive systems, including neurodivergent humans, artificial intelligences, and hybrid systems.

Users configure the engine with games, scenarios, objectives, characters, and environments, then run simulations through various clients (such as a web UI for human players).

Within a collaborative improvisational environment, player characters and other simulated characters interact as a cast with distinct abilities, goals, and cognitive profiles. Interaction unfolds primarily through language — including action, imagination, communication, and world-building — in a format inspired by tabletop roleplaying games.

How can I use it?

Try web demo instantly (no setup)

👉 Web Demo for a time and rate limited version of the engine. No setup required.

Quickstart

Alternatively, you can run the engine from your computer with your own API keys.

1. Install the package

pip install dcs-simulation-engine

# use --help after any dcs command to see usage info
dcs --help

2. Set up API keys

Add your API keys to a local .env file or set them as environment variables. See .env.example.

3. Run the engine

The first time you try and run it you will be prompted for your API keys if you haven't set them up in the .env file yet. You will also need to install Docker if you haven't already - the engine uses Docker containers to run the simulations.

dcs engine start --config examples/run_configs/demo.yml

Users can design, run, and deploy custom engine configurations—including their own games and characters.

👉 Usage for the full user guide.

Features

Item	Supported	Notes
Easy setup (pip, setup, run)	✅	Minimal friction onboarding
Out-of-the-box platform support	✅	Includes: built-in games & characters, default React web UI, autoplay client, local + Fly.io runs, reporting & analytics
Headless / modular usage	✅	Engine can run without default clients or deployment stack
Custom deployments & providers	✅	Containerized; deploy anywhere, plug in custom infra/UI
YAML run configurations	✅	Reproducible, config-driven runs
Configurable game parameters	✅	Optional game-specific configurations
Dev workflows & container	✅	Extensible + consistent onboarding
Example workflows	✅	Provided in examples/
Python game classes	✅	Flexible, expressive game logic
Game lifecycle validation	✅	Setup, step, finish, score
Python runtime performance	⚠️	Not ideal for ultra low-latency use cases; performant AI

See docs for more details on features, usage, and development.

---

Developed and maintained by the Diverse Cognitive Systems (DCS) group at Georgia Tech, within the Sonification Lab. Contact: [e-mail].