A terminal office for the work agents actually do

We are building theboringoffice as a native Go CLI because the place where coding agents work should not be another browser tab asking to be managed.

It is a startup office in your terminal. The boss is a real opencode session managed by Oikonomos. Employees are real opencode task sub-agents. The task board is backed by agentmemory actions; the mail room is backed by agentmemory signals. The floor gives those moving parts a shape you can understand at a glance.

Try the office before you trust it

Development tools should earn their place in a workflow. The fastest honest evaluation is a clearly labeled tour:

theboringoffice --demo

Demo mode simulates the same event flow the interface uses in live mode. It is designed to show the interaction without pretending there is a backend running behind it.

When you are ready to run a real shift, use the installer:

curl -fsSL \
  https://raw.githubusercontent.com/theboringhumane/theboringoffice/main/install.sh | sh

Then launch the office:

theboringoffice

Live mode starts opencode serve and opens the boss session. If you already have a server, attach directly instead:

theboringoffice --server http://127.0.0.1:4096

One terminal, six useful views

The right-side cockpit moves between chat, terminal, agents, board, mail, and activity with tab. These are not separate products. They are different views of the same shift:

• Chat is where you speak to the boss.
• Terminal is an embedded, real PTY shell.
• Agents shows the roster and current state of work.
• Board and mail carry agentmemory actions and signals.
• Activity makes the office's recent events visible.

Work threads keep sub-agent diffs, tool calls, and thinking attached to the task. Permission and question popovers surface decisions without trapping the rest of the workflow.

Context is part of the interface

The office normally restores the last chat on launch. To return to a particular past session, pass its ID:

theboringoffice --session <your-session-id>

The configuration lives at ~/.theboringoffice/configs/brain.json; inspect the defaults with theboringoffice --print-default-config.

This is a development-stage tool, and we are keeping the promise deliberately small: make real agent work visible, make human decisions manageable, and make useful context survive the next restart. If that changes how a solo developer or small team can use agents, the office has done its job.