orca

Product Introduction

1. Definition: Orca is an Agent Development Environment (ADE), a specialized desktop application designed to orchestrate multiple AI-powered coding agents (like Claude Code, Codex, and OpenCode) in parallel within isolated, git-based development environments called worktrees.
2. Core Value Proposition: It exists to solve the coordination and management overhead of running a fleet of AI coding assistants simultaneously. Orca provides the integrated tooling—terminals, file editor, browser, and git tracking—to automate and scale complex coding workflows, enabling developers to ship code faster by leveraging multiple AI agents in concert.

Main Features

1. Parallel, Isolated Worktrees: Each agent task runs in its own dedicated git worktree. This provides full filesystem isolation, preventing conflicts between concurrent AI agents. How it works: Orca leverages Git's worktree functionality to create separate working directories from a single repository clone, allowing parallel branches to be developed and tested independently without manual stashing or branch juggling.
2. Native Multi-Agent Terminal & Editor: Features a Ghostty-inspired terminal built with WebGL for performance, supporting infinite splits and restored scrollback. It includes a built-in VS Code-like file editor with autosave. This integrated environment allows agents to execute shell commands, edit code, and view results without context-switching between disparate tools.
3. Design Mode & Integrated Browser: Embeds a real Chromium window per worktree. Users can click any UI element to send its HTML, CSS, and a cropped screenshot directly into an agent's context. This enables AI agents to understand and modify live web applications visually, bridging the gap between code and its rendered output.
4. First-Party Git & Project Management Integration: Provides native GitHub and Linear integration. Users can browse PRs, issues, and project boards in-app, open a worktree from any task, review PRs, and create Linear issues with team selectors. Git operations like diff review, conflict resolution, and commit management are central to the workflow.
5. Orchestration & Automation CLI: Includes a full Command-Line Interface (CLI) that agents themselves can use to drive Orca (e.g., orca worktree create, snapshot, click). This enables the creation of automated, scripted workflows where agents can manage their own environments and tasks.
6. Remote SSH Worktrees & Mobile Companion: Supports secure SSH connections to remote development machines, providing full file editing, git, and terminal access with auto-reconnect. The companion iOS/Android app allows monitoring agent status, checking usage, switching AI accounts, and managing terminals remotely.

Problems Solved

1. Pain Point: Manual coordination and context switching between multiple AI coding agents and their outputs. Developers waste time managing separate terminal sessions, merging code from different branches, and reviewing disparate agent suggestions.
2. Target Audience: Software engineers, engineering teams, indie hackers, and DevOps professionals who actively use AI coding assistants like Claude Code or GitHub Copilot and want to scale their usage. Specifically, developers working on large refactors, parallel feature testing, or automated code maintenance.
3. Use Cases: Running A/B tests on code solutions by fanning one prompt to 5 different agents; maintaining a live agent to monitor and fix CI/CD pipeline failures; conducting parallel security audits or regression tests across different code branches; using the mobile app to keep long-running agents (e.g., tests, builds) progressing while away from the desk.

Unique Advantages

1. Differentiation: Unlike traditional IDEs (VS Code, Cursor) built for human-first interaction, Orca is an ADE built for agent-first orchestration. Unlike simple terminal wrappers that only run agents, Orca provides the complete surrounding environment (editor, browser, git, project management). It is cross-platform (macOS, Windows, Linux) and open-source (MIT licensed), unlike many macOS-only or closed-source alternatives.
2. Key Innovation: The deep integration of git worktrees as a first-class primitive for agent isolation. This technical approach, combined with a unified, agent-accessible interface for all development tools (terminal, editor, browser, git, CLI), creates a cohesive environment where AI agents can operate autonomously and in parallel at scale, which is not possible in standard IDEs.

Frequently Asked Questions (FAQ)

1. What is an Agent Development Environment (ADE)? An Agent Development Environment (ADE) is a specialized workspace designed to host, manage, and orchestrate multiple AI coding agents. Unlike an Integrated Development Environment (IDE) built for human developers, an ADE provides tooling for parallel execution, inter-agent coordination, and automated workflow management, treating AI agents as primary users.
2. How does Orca compare to Cursor? Cursor is an AI-enhanced IDE primarily designed for a single developer interacting with one AI assistant at a time. Orca is an ADE built to run and manage multiple AI agents (like Claude Code, Codex, OpenCode) simultaneously in parallel, isolated worktrees. Orca focuses on orchestration, fleet management, and providing agents with a full, automatable environment.
3. Can I use my existing Claude Code or Codex subscription with Orca? Yes, Orca is agent-agnostic. It works with your existing subscriptions for Claude Code, Codex, GitHub Copilot, OpenCode, Gemini, and any other AI coding agent that exposes a CLI. You configure your API keys or authentication, and Orca can run these agents side-by-side.
4. Is Orca free to use? Yes, Orca is completely free and open-source software released under the MIT License. You can download and use it on macOS, Windows, and Linux without cost.
5. What is the "Design Mode" in Orca? Design Mode is a feature that opens a real Chromium browser window within a worktree. You can click on any UI element, and Orca will capture its HTML structure, CSS styles, and a screenshot, then feed that technical and visual context directly into an AI agent's prompt. This is used for visually-guided frontend development and debugging.