Start playing

The terminal on this site is the game client. The scripted tier is live now: the mechanics are fully built, and the language model is coming in the next stage.

The game terminal

The entire game runs through text commands in a terminal. Open it and play a session.

Open the terminal

The scripted tier (S1) is live now: the engine, meters, decisions, effects and the retro all work fully. Standups and the retro run in templated mode — thinner prose, same mechanics. In stage S2 the language model connects, and characters will speak in the voice of their role.

Starting a session

1. Sign in

login

You will not need to type it again — the terminal remembers you.

2. Create a game

create <name> <scenario>

Replace <name> with anything and <scenario> with one of the five scenarios. A bare create Demo game works — it picks the default. See the Scenarios page for the full list.

3. Invite players

share

The command returns a link. Send it to your teammates — they appear at the same table. One to four players, bots fill empty seats.

What you will see in the terminal

The terminal is split into three layers.

Top row — team meters. TRUST, QUALITY, MORALE, SCOPE and others — each as a 10-segment bar with a numeric value.

Center — one of seven views. Switch with keys 17 or Tab:

KeyViewWhat it shows
1ActYour actions, current work-in-progress, hidden KPIs
2BoardTeam state at a glance: meters, risk telegraphs, policies, roster
3FeedSession event log
4StakeholdersPer-stakeholder detail
5ChatTeam chat
6TeamRoster, health radar, governance window
7RetroRetrospective — opens after the session ends

Bottom row — action panel. Your role’s available verbs with shortcut keys and a command line.

How a session goes

A scenario is split into rounds. Each round: events arrive → each seat sees its own slice → the team discusses → players spend capacity on actions → the engine resolves effects and moves the meters.

Decisions have three tiers: Solo (reversible, anytime), Quorum (needs teammate approval), Gate (irreversible, only at the standup).

The session ends with a retro: the engine compiles your decision chain, shows causal links, and states what everything cost.

What is next

Changelog — what is ready now and what is planned for stages S2 and S3.