The Terminal App
One screen for the whole game
ololo join and ololo start open a full-screen terminal app — the cockpit
for your session:
- Header — session name, phase, and the live countdown. Finish all your tasks early and it flips to "tasks done — waiting for session finish" while the others race.
- Sidebar — the current task with its description, the session leaderboard, and your recent check results. Open one to see the exact command, expected against actual output, and the points it moved.
- Agent terminal — your coding agent runs embedded in the app, in your working directory. It is a real terminal: type to your agent, interrupt it, or leave it alone, without leaving the game.
Hotkeys
| Key | Action |
|---|---|
F1 / ? | Help overlay with every hotkey |
F2 | Open the details of the last failed check |
F3 | Paste that failure into your agent's terminal |
F4 | Toggle the sidebar |
F9 | Switch focus between the game panels and the agent terminal |
F10 | Quit (when the game panels have focus) |
F3 is the power move: one key hands your agent exactly the check it should
fix next, with the command and the output it produced. Most of the game is
F2, F3, watch, repeat.
Picking an agent
Started without --agent? The app scans your PATH and shows a picker of the
agents it recognizes — see Supported Agents.
The app needs one of them to start. To play with no agent and write the code
yourself, use --no-tui instead: the checks score whatever is in your
directory, no matter who typed it.
Desktop and IDE agents — cursor, antigravity, zed — open in their own window. The game runs exactly the same, but the agent panel shows a placeholder rather than the conversation, because that conversation never passes through ololo.
Where things are stored
- Credentials:
~/.config/ololo/credentials.toml, one entry per profile. - Session log:
~/.config/ololo/PROFILE.tui.log(default.tui.logunless you use profiles) — attach it when you report a problem.