Getting work done in Rind: a CLI guide organized by scenario

These docs are not about architecture; they are only about how to get things done with the CLI. Each page maps to a real scenario: it starts from "I want to ...", not from a feature.

Architecture and implementation details live in internals; this is onboarding and daily usage.

Pick a scenario

What you want to do Read this page
Just installed, and want to watch it finish a task Get your first task done
The task is half done, and you want to change direction, queue, or interrupt Steer the work back on track
Connect a model, or switch between several models Get Rind to use the model you specify
A command will not finish soon, and you want it to keep going in the background Let unfinished commands keep going
Something was left unfinished, and you are coming back to continue Come back and keep working
Pin down project background and Skills Load the project into Rind
One person is not enough, so hand work to a team Hand the work to a team
Not sitting at the terminal; let scripts drive Rind Use Rind without being at the terminal
Forgot a key or a command Keyboard shortcuts and command cheatsheet

Build intuition in three minutes

  1. Start: rind runs inside the project directory; the session is only created when the first message is sent.
  2. Conversation: Enter sends; during a turn, Enter is steering, Tab queues.
  3. Interrupt: Ctrl+C interrupts the current turn; Ctrl+O expands tool output for details.
  4. Models: /login connects a provider, /model picks a model, Ctrl+T switches the reasoning effort.
  5. Background: console tools that exceed the wait window automatically move to the background; Ctrl+B shows progress.
  6. Scripts: rind run --prompt "..." gets the final answer; rind send --session <id> "..." injects into a running session.

Other entry points

The CLI is only one of Rind's Surfaces; the same core also has:

Back to the docs entry