zarlcode

zarlcode is a local, inspectable terminal coding-agent workbench. It runs in the workspace you launched it from, keeps model turns and tool calls visible in a TUI, and makes provider or model changes part of the normal workflow instead of a shell setup task.

Use it to investigate in read-only Plan mode, build with visible file and shell tools, inspect diffs, delegate focused sub-agents, and resume local sessions later.

Install

# Homebrew installs the binary as zarlcode
brew install zarldev/tap/zarlcode

From a source checkout:

go tool task zarlcode
# or
go run ./zarlcode/cmd

Note: go install github.com/zarldev/zarlmono/zarlcode/cmd@latest currently builds a binary named cmd because the CLI package directory is cmd. Use Homebrew, GitHub release archives, or go tool task zarlcode from a checkout until the public Go-install path is renamed or wrapped.

Start in the TUI

cd /path/to/your/workspace
zarlcode

On first run, press Enter to accept local defaults or Ctrl+S to configure a provider from settings. Sessions open in Build mode; press Shift+Tab to investigate in read-only Plan mode, then press it again when you want the agent to build. You do not need zarlcode init, keys, or a headless flag to begin.

• First run and onboarding — install, choose local defaults or configure a provider, and understand first-run vault setup.
• Your first workflow — Plan, Build, inspect diffs, and resume a local session.
• Interface guide — timeline, cockpit, working set, file viewer, model picker, settings, and keys.

What stays visible

The timeline shows model output, tool calls, command results, diffs, plans, and sub-agent summaries as they happen. The coding toolset reads files, makes anchored edits, searches the tree, and runs commands through a tracked process manager so long-running work can be inspected or stopped.

For larger tasks, zarlcode can compact older model context, delegate focused sub-agents, and resume prior sessions from local SQLite state. Its durable canonical transcript is separate from compactable provider history, so browsing and Markdown export retain the visible record after compaction.

• Sessions and transcripts explains persistence, resume, and export.
• Providers and credentials covers settings, models, OAuth, API keys, and the local credential vault.
• Safety and workspace access explains Plan mode, Build mode, shell policy, sandboxing, and external tools.

Build with confidence

zarlcode runs with your user’s privileges. Its workspace tools and shell commands are policy-guarded, and supported Linux systems can additionally sandbox shell execution, but a coding agent can still mutate files and run processes. Review plans, tool calls, and diffs when using powerful models or unfamiliar repositories.

For scripts, CI, credential subcommands, diagnostics, and one-shot headless work, use the separate automation and CLI reference. Those advanced surfaces remain supported without replacing the normal TUI journey.

Built on zkit

zarlcode is built with the same Go packages available in zkit:

Package	Role
zkit/agent/runner	The streaming agent loop.
zkit/ai/tools/code	Read, write, edit, shell, search, process, and plan tools.
zkit/agent/guardrails	Schema repair, shell policy, fan-out caps, and Go verifiers.
zkit/agent/coderunner	Standard coding toolset and guarded source assembly.
zkit/prefs and zkit/vault	Scoped settings and encrypted credential storage.

The source lives at zarlcode/.