Skip to content

app

Run many CLI (Coding) Agents in parallel lanes: issues, worktrees, sessions, and loops on one board. The agentic development environment.

Install

$/plugin marketplace add lanes-sh/app

Install command taken from the project's README or registry entry. Always review code before running it.

Lanes One workspace. Many agents. Run many AI coding agents in parallel lanes: issues, worktrees, sessions, one board.

Install ·
Quickstart ·
Docs ·
Discord

You lost track three agents ago

Five agents, eight terminal tabs. One is waiting for input. One finished ten minutes ago and nobody noticed. Two are quietly doing the same work.

Agents got good at writing code. Nothing got good at keeping track of them.

Install

brew install --cask lanes-sh/lanes/lanes && open -a Lanes

macOS Ventura or later. Universal binary, native on Apple Silicon and Intel. Lanes updates itself on launch.

The Quickstart takes you from here to a first agent session. Join the Discord for updates and questions.

What it is

A native macOS app that puts every agent session on an issue board. Each card is a task. Each task holds live agent terminals, its own git worktree, and the diff those agents produced. You see what is running, what is blocked, what is waiting on you, and what shipped, in one window.

Review it without leaving that window: the diff, the editor, the database, the commit, the pull request.

What you get

  • A board that holds the work. Drag issues through Planning, Implementation, Review and Done. Multi-select, labels, filters, dependencies with cycle detection, and board tabs scoped per project so two repos never bleed into each other.
  • Real terminals, several per issue. Every session is PTY-backed, and one card can hold the plan-mode session plus the two implementation runs it spawned. Live status, per-session token usage, and resume across restarts.
  • Claude Code and Codex, first class. Lanes detects them, reads their state, resumes their sessions and lists their models. Anything else runs as a plain shell session.
  • A worktree per issue. Generated branch names, uncommitted state in the status bar, setup and teardown scripts, and auto-cleanup when the issue completes.
  • A git client, not just a diff. Commit, publish, switch branches, undo the last unpushed commit, and open a pull request from the issue panel.
  • Monaco and a database browser. A real editor with a file tree, and read-only SQL against the SQLite files it finds in your project, so checking what an agent actually wrote is two clicks.
  • The model provider you choose. Gateway points a session at Ollama, LM Studio, vLLM, OpenRouter or z.ai, applied as real process environment so the request goes straight from the CLI to your provider. Local LLMs drives Ollama end to end. Both are research previews.
  • GitHub and Linear over OAuth. Import a whole sprint onto the board, and push results back as comments.
  • A team, when you want one. Remote workspaces sync issues, chat and presence in about two seconds, with assignees, presence avatars, and an honest banner when sync is degraded.

Full detail: Desktop docs.

Your apps, memory and skills

Lanes Link is one endpoint you run: Gmail, GitHub, Linear, Notion, Slack and over a hundred others, plus your memory, tasks and skills, behind one set of permissions.

Similar plugs

Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.

CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies

Documented system prompts from Anthropic - Claude Fable 5.1, Opus 5.5, Claude Design, Claude Code. OpenAI - ChatGPT GPT-6-Astra, Codex. Google - Gemini 3.8 Flash, 3.1 Pro,…

Wrap Antigravity, ChatGPT Codex, Claude Code, Grok Build, Muse Code, Devin as an OpenAI/Gemini/Claude/Codex compatible API service, allowing you to enjoy the free Gemini…