Help

Runyte is a terminal workspace for software development, built around a modal text editor. Normal mode is for selecting and navigating; Insert mode is for typing. Select what you want to change first, then apply an action. If that model is new to you, start with :tutorial—it opens a guided, hands-on introduction inside the editor.

Discover commands as you work

You do not need to memorize the interface before using it. Starting a command family such as Space, g, z, m, or Ctrl-w opens a hint popup with the available continuations and their descriptions. The : command palette is searchable and uses the same command registry.

  • Space ? opens contextual help for the current buffer type.
  • :help opens the general Runyte manual.
  • :help <topic> jumps to a topic such as git, search, mouse, or lsp.
  • :tutorial opens the interactive introduction.

Standalone and persistent workspaces

Runyte starts in standalone mode by default: the current process owns the editor workspace. In optional persistent mode, a local host keeps unsaved buffers, editor state, language servers, and terminal processes alive while the TUI client is detached.

runyte --persistent
runyte --session-list

Integrated terminal panes act as a workspace-scoped terminal multiplexer. You can split the interface, run shells or full-screen terminal applications, and review terminal history without stopping the child process.

Files and Git

Open a directory with runyte . to use the editable file explorer. Rename, move, copy, create, and delete entries with normal editor commands, then review an explicit filesystem plan before Runyte changes anything on disk.

The built-in Git interface covers status, diffs, staging, commits, branches, history, blame, stashes, and worktrees. Worktrees pair especially well with integrated terminals and coding agents: each task can have an isolated working directory while staying inside the same interface.

For the complete reference, see the Runyte user guide.