Orca

Orca is the free desktop for people who run several agents and want to see them at once. It supports 27 agents, Claude Code, Codex, Grok, Gemini, Cursor, GitHub Copilot, OpenCode, Amp, Pi, Hermes Agent and Goose among them, and the iOS and Android companions let you check a run from the sofa. The trait that decides it is that you can watch and nudge a run without being at the desk.

You need real parallel work, not curiosity, to justify the subscriptions underneath. Pick it if you have it and want MIT. Pick Agent Orchestrator if you want something to plan the work for you.

The risk is fan-out. The README's headline move is to fan one prompt across five agents, each in its own worktree, then compare and merge the winner, which means four of five runs are discarded and all five are billed to your subscriptions and rate limits. None of the five runs in a container; they share your machine and credentials.

Fan out for hard problems, not for every ticket. What it does right: AI diff annotation on the review pane, so the branch you keep comes with an explanation of what changed before you merge it.

Two documented mechanisms distinguish it. 1. Design Mode inspects UI in a real Chromium window, and clicking an element sends its HTML, CSS and a cropped screenshot straight into the agent's prompt, which is a precise way to gather visual context that most agents reconstruct from guesswork. 2. SSH remote worktrees run the agent on another machine with file editing, git and terminals, with auto-reconnect and port forwarding.

No benchmark is published, and the app claims none for itself. The observation: routing a screenshot into a prompt is a context-gathering step, and the quality of the edit still belongs to the agent underneath.

A Y Combinator company shipping a free MIT desktop app since March 2026 and collecting 60,881 stars is buying distribution, and the price of zero says the money is meant to come from something else in Stably's portfolio rather than from this window. Pricing power on the app itself: none, by design; the users are on other vendors' subscriptions.

Moat: the desktop habit, which is copyable. Likely acquirer: a lab that wants a neutral front end, or a terminal vendor. Likely pivot: a paid remote or team tier. Position: use it, expect a pricing page within the year.

The demo is five agents on one ticket. Procurement: the privacy documentation says anonymous usage data is collected and can be opted out of, which means it is on by default on sixty laptops until someone turns it off. There is no SSO, no audit log, no admin and no paid plan, so there is no counterparty; the Linear and GitHub integrations use each engineer's own credentials.

Builds exist for macOS, Windows and Linux, so nobody is excluded. It does not run in CI. Onboarding is a cask. Approved with conditions: telemetry off in a managed config, personal use only.

MIT, and packaged properly: a Homebrew cask and stably-orca-bin on the AUR, which is more Linux respect than most Electron-era apps show. The CLI is the part I use: orca worktree create, snapshot, click and fill mean a run is a shell script, not a sequence of mouse gestures, and the app becomes something my own tools can drive.

No MCP and no model configuration, because the agents in the worktrees bring their keys and their local endpoints. Forking is a big desktop project, not a weekend. Open enough, scriptable enough, and I did not expect either.