Documentation

Install stepper via prebuilt binaries, build from source, and configure your first run with guided setup or environment overrides.

Get started with stepper and reference the CLI commands, flags, and modes.

Configure stepper with camelCase JSON: user-level base merged with project settings, featuring providers, layers, permissions, and MCP servers.

Configure API keys for five AI providers with environment variables, OS keyring, or explicit config.

Configure per-layer models, permissions, tools, and skills; run layers sequentially or in parallel fan-out with task assignment.

Permission modes and rule specifiers for controlling agent tool access.

Built-in slash commands, custom commands with permission-gated substitutions, skills with progressive disclosure, and MCP server integration.

Dispatch, compaction, checkpoints, sessions, hooks, interrupts, and prompt caching features for layered AI coding workflows.

Automated CI pipelines, manual workflow runs, and cross-platform release binaries via GitHub Actions.

10-crate Cargo workspace design: crate roles, isolation invariants, and data flow from CLI bootstrap through multi-layer agent orchestration to TUI.

What stepper has shipped, what is being hardened for real-world use, and the limits deliberately deferred.