Skip to main content
Back to timeline
stablyai/orcaSource publication:

Orca runs coding agents in parallel, each in its own worktree, with chat, terminal and diffs in one app

Synopsis

Stablyai released Orca, an open-source app that runs command-line coding agents such as Claude Code and Codex in parallel, giving each agent its own git worktree. Chat and terminal, diff review, a browser, a CLI and phone pairing sit in the same application, which also accepts files or images dragged into an agent's prompt, schedules prompts, and lets agents drive desktop apps when no CLI or API exists. According to the release notes, any CLI agent that runs in a terminal can connect to Orca.

AI-generated editorial illustration: stablyai / orca

Interpretation

Coding agents can work in parallel, each in its own git worktree, so edits made at the same time stay in separate places. The release notes present this parallel-worktree arrangement as a core capability and state that Orca works with any CLI agent that runs in a terminal. The release notes describe the arrangement as running Claude Code, Codex and any other coding agent in parallel, each in its own worktree, and list a long roster of CLI agents that can connect.

The interface keeps an agent's changes and outputs beside it: chat or terminal, diffs, a browser and a CLI all sit in one application window. The release notes make this unified interface the product's positioning, saying chat or terminal, diffs, a browser and a CLI are in one app. The release notes list concrete features, including dragging files or images into a prompt, a single shortcut that jumps across repos and hosts, and previews of images, PDFs and Markdown inside the workspace.

Agents can also drive desktop apps and external browsers that have no CLI or API, and can tap, type and check their own work in an iOS Simulator or Android emulator. The release notes describe Computer Use as letting agents see and drive desktop apps and external browsers when there is no CLI or API. The release notes list Computer Use and Mobile Emulator among the features shipped with the app and say agents can tap, type and check their own work in the emulator.

Perspective

The notes address developers who use command-line coding agents and cover parallel worktrees, a unified interface, file dragging, scheduled prompts, desktop control and a mobile emulator; they describe what the app offers, not how those capabilities perform on specific tasks.

The release notes say the feature list perpetually lags daily shipping and that the changelog is the real feature list, so the listed capabilities may not be the complete set in the current version; the notes also give no performance data for these features on specific tasks.