[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f11w0vrbqfuik":3},{"code":4,"message":5,"data":6,"timestamp":52},10000,"success",{"full_name":7,"position":8,"description":9,"language":10,"stars":11,"forks":12,"stars_gained":13,"age_days":14,"hn_threads":15,"one_liner":22,"summary_md":23,"topic":24,"github_topics":27,"license":48,"homepage":49,"updated_at":50,"history":51},"stablyai\u002Forca",13,"Orca is the place for working with a fleet of agents (e.g. Grok Bot, Claude Code). Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.","TypeScript",104527,6603,27848,221,[16],{"id":17,"title":18,"points":19,"comments":20,"posted_at":21},47549197,"Show HN: Open Source 'Conductor + Ghostty'",21,10,"2026-03-27","An orchestration desktop for running many coding agents in parallel git worktrees, with phone monitoring and remote execution.","## The problem\n\nOnce you run more than one coding agent (Claude Code, Codex, OpenCode, Pi), the bottleneck stops being the model and becomes logistics: which agent is touching which branch, what did it change, and how do you compare two attempts without clobbering your working tree. Most people juggle terminal tabs and copy results between checkouts by hand. Orca puts each agent run in its own git worktree and gives you one window to launch, watch, review and merge those runs.\n\n## What's different\n\nThe usual approach is a plain terminal multiplexer plus manual `git worktree` commands, or a single-agent IDE extension that assumes one conversation at a time. Orca's bet is the fleet as the unit of work: fan one prompt across several agents in isolated worktrees, then compare diffs and keep the winner. Two things genuinely set it apart rather than pad the list. First, the remote and mobile surfaces — worktrees can run over SSH on a bigger box, and a companion app lets you nudge or approve an agent from your phone, which is the difference between watching a long run and being chained to your desk. Second, the CLI exposes the app itself (`orca worktree create`, `snapshot`, `click`), so agents can script Orca rather than only being driven by it.\n\nThe trade-off is surface area: Orca is trying to be terminal, editor, browser, GitHub\u002FLinear client and orchestrator in one binary, which is a lot of trust to place in one vendor's opinion of your workflow.\n\n## Who it's for\n\nDevelopers who already pay for one or more agent subscriptions and routinely run several attempts on the same task — reviewers comparing candidate diffs, people working across many repos, and anyone whose local machine can't host the agents they want to run. Solo developers with a single agent and a single repo will get little from the worktree machinery and should stay with their terminal and an editor extension. Teams that need auditable, org-wide agent governance should look at platform-level offerings instead; Orca is a personal cockpit.\n\n## Cost of adoption\n\nOrca is a desktop application (macOS, Windows, Linux) with MIT-licensed source, plus a mobile companion on the App Store and as an Android APK. There's no install command in the README — you download a build. Real costs are workflow-shaped, not financial: you adopt a new IDE, your terminals and diffs move inside it, and your agents inherit whatever account and usage-tracking setup Orca manages. The README is long on features and short on internals — no architecture notes, no statement about where prompts or code travel beyond what the vendor's own docs say. Treat the SSH and remote runtime paths as the ones to validate first, since that's where credentials and code leave your machine.\n\n## Verdict\n\nThe repository was created in March 2026 and has already crossed 100k stars, with roughly 9.5k added in a single week in late October — that is a project on a steep spike, not a decade-old staple, and the star count should be read as attention rather than proven durability. The idea is sound and the worktree-per-agent model is the right primitive; the Hacker News interest was modest, so the growth is coming from somewhere other than one viral thread. Watch release cadence and issue response for a few months before making it your only cockpit. Adopt it if parallel agents are already your daily reality, not if you're merely curious.",{"slug":25,"name":26},"developer-tools","Developer Tools",[28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47],"ade","agent-ide","ai-agents","automation","bot","claude-code","codex","ghostty","grok-bot","hermes-agent","mobile-app","muse","openclaw","opencode","orchestration","pi","runtime","terminal","worktrees","yc-backed","MIT","https:\u002F\u002Forca.dev","2026-10-24T14:10:12Z",[],1792864187]