[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"public-config":3,"i-flag:ph-4x3":14,"i-flag:us-4x3":19,"i-flag:de-4x3":21,"i-flag:hk-4x3":23,"i-tabler:lifebuoy":25,"i-tabler:timeline":28,"i-tabler:apps":30,"i-tabler:package":32,"i-tabler:circle-half-2":34,"i-tabler:sun":36,"i-tabler:moon":38,"i-tabler:menu-2":40,"i-tabler:user":42,"i-tabler:notebook":44,"i-simple-icons:forgejo":46,"i-tabler:shield-lock":48,"i-tabler:tools":50,"i-tabler:calendar":52,"i-tabler:x":54,"$f2gekvcmsv2gz0":56,"i-tabler:arrow-left":74,"i-tabler:tag":76},{"matomo":4,"booking":7,"site":9,"blog":11,"newsletter":12},{"enabled":5,"siteId":6},true,"7",{"enabled":5,"url":8},"https:\u002F\u002Fcal.com\u002Fjdbungart",{"url":10},"https:\u002F\u002Fjdneer.com",{"enabled":5},{"enabled":13},false,{"left":15,"top":15,"width":16,"height":17,"rotate":15,"vFlip":13,"hFlip":13,"body":18},0,640,480,"\u003Cpath fill=\"#0038a8\" d=\"M0 0h640v240H0z\"\u002F>\u003Cpath fill=\"#ce1126\" d=\"M0 240h640v240H0z\"\u002F>\u003Cpath fill=\"#fff\" d=\"M415.7 240L0 480V0\"\u002F>\u003Cpath fill=\"#fcd116\" d=\"M26.7 42.4L41 55l16.6-9.2l-7.4 17.5l14 13l-19-1.6l-8.1 17.2l-4.3-18.5L14 71l16.3-10zm323.8 172.3l.4 19l18 6.3l-18 6.2l-.4 19l-11.5-15.1l-18.2 5.5l10.8-15.6l-10.8-15.6l18.2 5.5zM37.2 388.1l8 17.2l19-1.6l-13.9 13l7.4 17.5l-16.6-9.1l-14.4 12.4l3.6-18.7L14 409l18.9-2.4zm114.2-249l-6.2 6.2l3.1 47l-3 .3l-5.7-42.9l-5.1 5l7.6 38.4a48 48 0 0 0-17.2 7.1l-21.7-32.4H96l26.4 34.3l-2.4 2l-31.1-35.5h-8.8v8.8l35.4 31l-2 2.5l-34.3-26.3v7.1l32.5 21.7q-5.2 7.8-7.1 17.2L66.3 223l-5.1 5l42.9 5.7q-.3 1.6-.3 3.1l-47-3l-6.2 6.2l6.2 6.2l47-3.1l.3 3.1l-42.9 5.7l5 5l38.4-7.6a48 48 0 0 0 7.1 17.2l-32.5 21.7v7.2l34.3-26.3l2 2.4l-35.4 31v8.8H89l31-35.4l2.5 2L96 312.2h7.2l21.7-32.5q7.8 5.2 17.2 7.1l-7.6 38.4l5 5l5.7-42.9q1.5.3 3.1.3l-3 47l6.1 6.2l6.3-6.2l-3.1-47l3-.3l5.7 43l5.1-5.1l-7.6-38.4a48 48 0 0 0 17.2-7.1l21.7 32.5h7.2l-26.4-34.3l2.4-2l31.1 35.4h8.8v-8.8l-35.4-31l2-2.4l34.3 26.3v-7.2l-32.5-21.7q5.2-7.8 7.1-17.2l38.3 7.6l5.1-5l-42.9-5.7q.3-1.5.3-3.1l47 3l6.2-6.1l-6.2-6.2l-47 3l-.3-3l42.9-5.7l-5-5l-38.4 7.5a48 48 0 0 0-7.1-17.2l32.5-21.7v-7.1l-34.3 26.3l-2-2.4l35.4-31v-8.9H214l-31 35.5l-2.5-2l26.4-34.3h-7.2L178 200.2q-7.8-5.2-17.2-7.1l7.6-38.3l-5-5l-5.7 42.8l-3.1-.3l3-47z\"\u002F>",{"left":15,"top":15,"width":16,"height":17,"rotate":15,"vFlip":13,"hFlip":13,"body":20},"\u003Cpath fill=\"#bd3d44\" d=\"M0 0h640v480H0\"\u002F>\u003Cpath stroke=\"#fff\" stroke-width=\"37\" d=\"M0 55.3h640M0 129h640M0 203h640M0 277h640M0 351h640M0 425h640\"\u002F>\u003Cpath fill=\"#192f5d\" d=\"M0 0h364.8v258.5H0\"\u002F>\u003Cmarker id=\"SVGEq3dreKJ\" markerHeight=\"30\" markerWidth=\"30\">\u003Cpath fill=\"#fff\" d=\"m14 0l9 27L0 10h28L5 27z\"\u002F>\u003C\u002Fmarker>\u003Cpath fill=\"none\" marker-mid=\"url(#SVGEq3dreKJ)\" d=\"m0 0l16 11h61h61h61h61h60L47 37h61h61h60h61L16 63h61h61h61h61h60L47 89h61h61h60h61L16 115h61h61h61h61h60L47 141h61h61h60h61L16 166h61h61h61h61h60L47 192h61h61h60h61L16 218h61h61h61h61h60z\"\u002F>",{"left":15,"top":15,"width":16,"height":17,"rotate":15,"vFlip":13,"hFlip":13,"body":22},"\u003Cpath fill=\"#fc0\" d=\"M0 320h640v160H0z\"\u002F>\u003Cpath fill=\"#000001\" d=\"M0 0h640v160H0z\"\u002F>\u003Cpath fill=\"red\" d=\"M0 160h640v160H0z\"\u002F>",{"left":15,"top":15,"width":16,"height":17,"rotate":15,"vFlip":13,"hFlip":13,"body":24},"\u003Cpath fill=\"#ec1b2e\" d=\"M0 0h640v480H0\"\u002F>\u003Cpath id=\"SVG5tfvBPpj\" fill=\"#fff\" d=\"M346.3 103.1C267 98 230.6 201.9 305.6 240.3c-26-22.4-20.6-55.3-10.1-72.4l1.9 1.1c-13.8 23.5-11.2 52.7 11.1 71c-12.7-12.3-9.5-39 12.1-48.9s23.6-39.3 16.4-49.1q-14.7-25.6 9.3-38.9M307.9 164l-4.7 7.4l-1.8-8.6l-8.6-2.3l7.8-4.3l-.6-8.9l6.5 6.1l8.3-3.3l-3.7 8.1l5.6 6.8z\"\u002F>\u003Cuse href=\"#SVG5tfvBPpj\" transform=\"rotate(72 312.5 243.5)\"\u002F>\u003Cuse href=\"#SVG5tfvBPpj\" transform=\"rotate(144 312.5 243.5)\"\u002F>\u003Cuse href=\"#SVG5tfvBPpj\" transform=\"rotate(216 312.5 243.5)\"\u002F>\u003Cuse href=\"#SVG5tfvBPpj\" transform=\"rotate(288 312.5 243.5)\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":27},24,"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M8 12a4 4 0 1 0 8 0a4 4 0 1 0-8 0\"\u002F>\u003Cpath d=\"M3 12a9 9 0 1 0 18 0a9 9 0 1 0-18 0m12 3l3.35 3.35M9 15l-3.35 3.35m0-12.7L9 9m9.35-3.35L15 9\"\u002F>\u003C\u002Fg>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":29},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"m4 16l6-7l5 5l5-6\"\u002F>\u003Cpath d=\"M14 14a1 1 0 1 0 2 0a1 1 0 1 0-2 0M9 9a1 1 0 1 0 2 0a1 1 0 1 0-2 0m-6 7a1 1 0 1 0 2 0a1 1 0 1 0-2 0m16-8a1 1 0 1 0 2 0a1 1 0 1 0-2 0\"\u002F>\u003C\u002Fg>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":31},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M4 5a1 1 0 0 1 1-1h4a1 1 0 0 1 1 1v4a1 1 0 0 1-1 1H5a1 1 0 0 1-1-1zm0 10a1 1 0 0 1 1-1h4a1 1 0 0 1 1 1v4a1 1 0 0 1-1 1H5a1 1 0 0 1-1-1zm10 0a1 1 0 0 1 1-1h4a1 1 0 0 1 1 1v4a1 1 0 0 1-1 1h-4a1 1 0 0 1-1-1zm0-8h6m-3-3v6\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":33},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"m12 3l8 4.5v9L12 21l-8-4.5v-9zm0 9l8-4.5M12 12v9m0-9L4 7.5m12-2.25l-8 4.5\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":35},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M3 12a9 9 0 1 0 18 0a9 9 0 1 0-18 0m9-9v18m0-7l7-7m-7 12l8.5-8.5M12 9l4.5-4.5\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":37},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M8 12a4 4 0 1 0 8 0a4 4 0 1 0-8 0m-5 0h1m8-9v1m8 8h1m-9 8v1M5.6 5.6l.7.7m12.1-.7l-.7.7m0 11.4l.7.7m-12.1-.7l-.7.7\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":39},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M12 3h.393a7.5 7.5 0 0 0 7.92 12.446A9 9 0 1 1 12 2.992z\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":41},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M4 6h16M4 12h16M4 18h16\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":43},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M8 7a4 4 0 1 0 8 0a4 4 0 0 0-8 0M6 21v-2a4 4 0 0 1 4-4h4a4 4 0 0 1 4 4v2\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":45},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M6 4h11a2 2 0 0 1 2 2v12a2 2 0 0 1-2 2H6a1 1 0 0 1-1-1V5a1 1 0 0 1 1-1m3 0v18m4-14h2m-2 4h2\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":47},"\u003Cpath fill=\"currentColor\" d=\"M16.777 0a2.9 2.9 0 1 1-2.529 4.322H12.91a4.266 4.266 0 0 0-4.265 4.195v2.118a7.08 7.08 0 0 1 4.147-1.42l.118-.002h1.338a2.9 2.9 0 0 1 5.43 1.422a2.9 2.9 0 0 1-5.43 1.422H12.91a4.266 4.266 0 0 0-4.265 4.195v2.319A2.9 2.9 0 0 1 7.222 24A2.9 2.9 0 0 1 5.8 18.57V8.589a7.11 7.11 0 0 1 6.991-7.108l.118-.001h1.338A2.9 2.9 0 0 1 16.778 0M7.223 19.905a1.194 1.194 0 1 0 0 2.389a1.194 1.194 0 0 0 0-2.389m9.554-10.464a1.194 1.194 0 1 0 0 2.389a1.194 1.194 0 0 0 0-2.39m0-7.735a1.194 1.194 0 1 0 0 2.389a1.194 1.194 0 0 0 0-2.389\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":49},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M12 3a12 12 0 0 0 8.5 3A12 12 0 0 1 12 21A12 12 0 0 1 3.5 6A12 12 0 0 0 12 3\"\u002F>\u003Cpath d=\"M11 11a1 1 0 1 0 2 0a1 1 0 1 0-2 0m1 1v2.5\"\u002F>\u003C\u002Fg>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":51},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M3 21h4L20 8a1.5 1.5 0 0 0-4-4L3 17zM14.5 5.5l4 4\"\u002F>\u003Cpath d=\"M12 8L7 3L3 7l5 5M7 8L5.5 9.5M16 12l5 5l-4 4l-5-5m4 1l-1.5 1.5\"\u002F>\u003C\u002Fg>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":53},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M4 7a2 2 0 0 1 2-2h12a2 2 0 0 1 2 2v12a2 2 0 0 1-2 2H6a2 2 0 0 1-2-2zm12-4v4M8 3v4m-4 4h16m-9 4h1m0 0v3\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":55},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M18 6L6 18M6 6l12 12\"\u002F>",{"slug":57,"title":58,"excerpt":59,"image":60,"publishedAt":61,"topics":62,"readMinutes":63,"html":64,"metaTitle":58,"metaDescription":65,"shareImage":66,"updatedAt":61,"author":70,"related":73},"orca-wants-to-be-the-workbench-for-your-fleet-of-coding-agents-heres-what-developers-say","Orca Wants to Be the Workbench for Your Fleet of Coding Agents. Here's What Developers Say","Orca is a free, MIT-licensed desktop app from Stably AI that runs Claude Code, Codex and other coding agents side by side, each in its own git worktree. Developers who use it like the workflow, but warn about token costs, review load and one security gap that is now fixed.","\u002Fcontent\u002Fimages\u002F2026\u002F10\u002Forcas-stephen-walker-unsplash.jpg","2026-10-18T22:45:47.000+08:00",[],7,"\u003Cp>Orca is a desktop app built around a simple idea: if you already run AI coding agents, you should be able to run several of them at once without them tripping over each other. Its maker, Stably AI, calls it an \"agent development environment,\" or ADE. Every task gets its own git worktree, its own agent terminal and its own browser tab, so Claude Code can fix a bug in one worktree while Codex tries a migration in another. The \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002F?ref=ghost.jdneer.com\">project site\u003C\u002Fa> says it works with Claude Code, Codex, Grok, Gemini, Cursor, GitHub Copilot, OpenCode and \"any other CLI agent.\"\u003C\u002Fp>\u003Cp>Orca does not ship a model of its own. According to its \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fdocs?ref=ghost.jdneer.com\">documentation\u003C\u002Fa>, it runs the agents and subscriptions a developer already has, and it is aimed at \"people who already write code for a living.\" The docs are blunt about it: \"If you're looking for a no-code tool, Orca is not that.\"\u003C\u002Fp>\n\n\u003Ctable>\u003Ctbody>\u003Ctr>\u003Ctd>\u003Cstrong>Maker\u003C\u002Fstrong>\u003C\u002Ftd>\u003Ctd>Stably AI, San Francisco (Y Combinator backed)\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Cstrong>License and price\u003C\u002Fstrong>\u003C\u002Ftd>\u003Ctd>MIT, free; agents billed by their own providers\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Cstrong>Desktop\u003C\u002Fstrong>\u003C\u002Ftd>\u003Ctd>macOS, Windows, Linux (AppImage, .deb, .rpm)\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Cstrong>Mobile\u003C\u002Fstrong>\u003C\u002Ftd>\u003Ctd>iOS app, Android APK (companion only)\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Cstrong>Latest release\u003C\u002Fstrong>\u003C\u002Ftd>\u003Ctd>v1.4.229, October 17, 2026\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>\u003Cstrong>Source\u003C\u002Fstrong>\u003C\u002Ftd>\u003Ctd>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca?ref=ghost.jdneer.com\">github.com\u002Fstablyai\u002Forca\u003C\u002Fa>\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\n\n\u003Cfigure class=\"kg-card kg-image-card kg-card-hascaption\">\u003Cimg src=\"\u002Fcontent\u002Fimages\u002F2026\u002F10\u002Forca-readme-hero.jpg\" class=\"kg-image\" alt=\"Orca desktop app running Claude Code and Codex in parallel worktrees, with the mobile companion app\" loading=\"lazy\" width=\"1998\" height=\"1250\" srcset=\"\u002Fcontent\u002Fimages\u002Fsize\u002Fw600\u002F2026\u002F10\u002Forca-readme-hero.jpg 600w, \u002Fcontent\u002Fimages\u002Fsize\u002Fw1000\u002F2026\u002F10\u002Forca-readme-hero.jpg 1000w, \u002Fcontent\u002Fimages\u002Fsize\u002Fw1600\u002F2026\u002F10\u002Forca-readme-hero.jpg 1600w, \u002Fcontent\u002Fimages\u002F2026\u002F10\u002Forca-readme-hero.jpg 1998w\" sizes=\"(min-width: 720px) 720px\" \u002F>\u003Cfigcaption>Orca running Claude Code and Codex side by side in separate worktrees, with the mobile companion app in front. Source: \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca?ref=ghost.jdneer.com\">Orca README on GitHub\u003C\u002Fa> (MIT license).\u003C\u002Ffigcaption>\u003C\u002Ffigure>\u003Ch2 id=\"what-it-does\">What it does\u003C\u002Fh2>\u003Cp>Beyond parallel worktrees, the \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca?ref=ghost.jdneer.com\">README\u003C\u002Fa> lists a diff viewer where developers can leave comments on specific lines and send them back to the agent, an embedded Chromium browser with a \"Design Mode\" that sends a clicked element's HTML, CSS and a screenshot into the agent's prompt, SSH worktrees for running agents on a remote machine, built-in GitHub and Linear views, and usage tracking for Claude and Codex accounts. An \u003Ccode>orca\u003C\u002Fcode> command line tool lets agents drive Orca themselves, for example to create worktrees. A mobile companion app shows agent status and lets users send follow-ups from a phone, but it pairs with a running desktop or server and does not work on its own.\u003C\u002Fp>\u003Cfigure class=\"kg-card kg-image-card kg-card-hascaption\">\u003Cimg src=\"\u002Fcontent\u002Fimages\u002F2026\u002F10\u002Forca-annotate-ai-diff.jpg\" class=\"kg-image\" alt=\"Orca diff viewer showing changed lines in a markdown file next to an agent terminal\" loading=\"lazy\" width=\"960\" height=\"542\" srcset=\"\u002Fcontent\u002Fimages\u002Fsize\u002Fw600\u002F2026\u002F10\u002Forca-annotate-ai-diff.jpg 600w, \u002Fcontent\u002Fimages\u002F2026\u002F10\u002Forca-annotate-ai-diff.jpg 960w\" sizes=\"(min-width: 720px) 720px\" \u002F>\u003Cfigcaption>The diff viewer, where developers can comment on specific lines and send the notes back to the agent. Source: \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fdocs\u002Freview\u002Fannotate-ai-diff?ref=ghost.jdneer.com\">Orca docs\u003C\u002Fa> (MIT-licensed repo asset).\u003C\u002Ffigcaption>\u003C\u002Ffigure>\u003Cp>The project moves fast. The GitHub repository was created in March 2026, has passed 96,000 stars, and ships desktop releases almost daily. The newest on GitHub is \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca\u002Freleases?ref=ghost.jdneer.com\">v1.4.229\u003C\u002Fa>, published October 17. The \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fchangelog?ref=ghost.jdneer.com\">changelog on the website\u003C\u002Fa> runs a little behind the GitHub releases page; the README itself calls the GitHub release notes \"the real feature list.\"\u003C\u002Fp>\u003Ch2 id=\"what-developers-report\">What developers report\u003C\u002Fh2>\u003Cp>Orca has not drawn coverage from major tech publications yet, so practical feedback comes mostly from individual developers writing about their own use, plus discussion on GitHub and Hacker News. The testimonials on Orca's homepage are posts the company picked itself, so they are left out here.\u003C\u002Fp>\u003Cp>\u003Cstrong>The worktree model is the main draw.\u003C\u002Fstrong> Developer Angélo Lima, who \u003Ca href=\"https:\u002F\u002Fangelo-lima.fr\u002Fen\u002Forca-ide-agentic-development-en\u002F?ref=ghost.jdneer.com\">wrote up six weeks\u003C\u002Fa> of daily use on October 15, says one worktree per session changed how he works because he no longer has to run tasks one after another. \"Code doesn't come out of any single session faster, but nothing waits for the previous one to finish anymore,\" he writes. He also leans on Orca's scheduled automations for routine jobs such as a weekly pass over Dependabot pull requests.\u003C\u002Fp>\u003Cp>\u003Cstrong>The phone app gets real use.\u003C\u002Fstrong> James Carr, an indie developer who \u003Ca href=\"https:\u002F\u002Fjames-carr.org\u002Fposts\u002F2026-09-30-current-agentic-development-setup-orca-oh-my-pi\u002F?ref=ghost.jdneer.com\">moved to Orca after close to two years on Cursor\u003C\u002Fa>, pairs it with the Oh My Pi agent harness. \"The Orca feature I love most is relay, which lets me run sessions from my phone,\" he writes, describing kicking off work from his phone and hearing pull requests open later.\u003C\u002Fp>\u003Cp>\u003Cstrong>It does not lower the bill.\u003C\u002Fstrong> Several reviewers raise the same point. Carr notes that \"Orca doesn’t make a dent in the model bill, but it lets many worktrees run at once.\" Lima puts it more directly: \"The tool makes parallelism easy. It doesn't make it free.\" Full-stack engineer Ramón Chancay, in a \u003Ca href=\"https:\u002F\u002Fwww.ramonchancay.me\u002Fblog\u002Forca-parallel-coding-agent-fleet?ref=ghost.jdneer.com\">July review\u003C\u002Fa>, warns that running one task across several agents multiplies token use, and that with many worktrees open, human review \"becomes the bottleneck.\"\u003C\u002Fp>\u003Cp>\u003Cstrong>Resources and rough edges.\u003C\u002Fstrong> Lima says each worktree carries its own dependencies, such as \u003Ccode>node_modules\u003C\u002Fcode>, so disk space fills quickly on large projects. He also lists the near-daily release pace as a stability trade-off and points out that \"Orca doesn't support VS Code extensions,\" so he still opens a traditional IDE for heavy manual refactoring with a debugger.\u003C\u002Fp>\u003Cp>\u003Cstrong>Linux support stood out early.\u003C\u002Fstrong> When the team first posted Orca as a \u003Ca href=\"https:\u002F\u002Fnews.ycombinator.com\u002Fitem?id=47549197&ref=ghost.jdneer.com\">Show HN\u003C\u002Fa> in March, one commenter \u003Ca href=\"https:\u002F\u002Fnews.ycombinator.com\u002Fitem?id=47549215&ref=ghost.jdneer.com\">wrote\u003C\u002Fa>: \"It's been working super well for me. First tool that actually works well on Linux.\" The Hacker News thread was small, with about 10 comments.\u003C\u002Fp>\u003Ch2 id=\"a-security-concern-and-the-fix\">A security concern, and the fix\u003C\u002Fh2>\u003Cp>The most pointed criticism came from developer Martin Tapia, who \u003Ca href=\"https:\u002F\u002Fmartintapia.com\u002Fblog\u002F2026\u002Fi-tried-orca-and-vscode-agents-window-i-went-back-to-tmux-shortly-after?ref=ghost.jdneer.com\">wrote on September 30\u003C\u002Fa> that he ran agents in sandboxed VMs connected to Orca over SSH, then found that the \u003Ccode>orca\u003C\u002Fcode> command inside a sandbox could list and send commands to any terminal open in his Orca app, including local ones and other SSH hosts. That defeats the point of a sandbox if an agent is hit by a prompt injection. He filed \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca\u002Fissues\u002F23835?ref=ghost.jdneer.com\">GitHub issue #23835\u003C\u002Fa>, and maintainer nwparker replied: \"It's a valid concern. I think it's a give and take as always.\" Tapia's verdict on his blog: \"So I uninstalled Orca.\"\u003C\u002Fp>\u003Cp>Stably did fix it. A pull request titled \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca\u002Fpull\u002F26337?ref=ghost.jdneer.com\">\"deny-by-default caller permissions and a scoped SSH bridge credential\"\u003C\u002Fa> closed the issue on October 8, and the \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca\u002Freleases\u002Ftag\u002Fv1.4.224?ref=ghost.jdneer.com\">v1.4.224 release notes\u003C\u002Fa> (October 10) say \u003Ccode>orca\u003C\u002Fcode> commands on an SSH host now reach only that host's own terminals by default. Broader control is an opt-in per host, under Settings, SSH Hosts, Advanced Connection. Anyone who isolates agents this way should run v1.4.224 or later and leave that switch off for untrusted hosts.\u003C\u002Fp>\u003Ch2 id=\"running-orca-with-local-models\">Running Orca with local models\u003C\u002Fh2>\u003Cp>Orca has no model or provider settings of its own. Its \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fdocs\u002Fagents\u002Fsupported?ref=ghost.jdneer.com\">supported agents page\u003C\u002Fa> says it plainly: \"the agent combobox just launches a process in a terminal.\" Running Orca on local models therefore means pointing the coding agent at a local inference server, then launching that agent from Orca's picker as usual. Orca's Settings, Agents page can hold a custom command, default arguments and environment variables for each agent, which is a handy place to keep a local setup apart from a cloud one.\u003C\u002Fp>\u003Cp>How the agent connects depends on the agent. \u003Cstrong>Claude Code\u003C\u002Fstrong> reads \u003Ccode>ANTHROPIC_BASE_URL\u003C\u002Fcode>, so it works with servers that speak Anthropic's Messages API, including Ollama, LM Studio and llama.cpp's \u003Ccode>llama-server\u003C\u002Fcode>. Ollama's \u003Ca href=\"https:\u002F\u002Fdocs.ollama.com\u002Fintegrations\u002Fclaude-code?ref=ghost.jdneer.com\">Claude Code guide\u003C\u002Fa> uses:\u003C\u002Fp>\u003Cpre>\u003Ccode>export ANTHROPIC_AUTH_TOKEN=ollama\nexport ANTHROPIC_API_KEY=\"\"\nexport ANTHROPIC_BASE_URL=http:\u002F\u002Flocalhost:11434\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>\u003Cstrong>Codex\u003C\u002Fstrong> has a built-in local mode. Per OpenAI's \u003Ca href=\"https:\u002F\u002Fdevelopers.openai.com\u002Fcodex\u002Fconfig-advanced?ref=ghost.jdneer.com\">configuration docs\u003C\u002Fa>, \u003Ccode>codex --oss\u003C\u002Fcode> runs against Ollama or LM Studio, and \u003Ccode>oss_provider = \"ollama\"\u003C\u002Fcode> (or \u003Ccode>\"lmstudio\"\u003C\u002Fcode>) in \u003Ccode>config.toml\u003C\u002Fcode> sets the default. \u003Cstrong>OpenCode\u003C\u002Fstrong> takes any OpenAI-compatible endpoint: its \u003Ca href=\"https:\u002F\u002Fopencode.ai\u002Fdocs\u002Fproviders\u002F?ref=ghost.jdneer.com\">providers docs\u003C\u002Fa> show a provider block in \u003Ccode>opencode.json\u003C\u002Fcode> with \u003Ccode>\"npm\": \"@ai-sdk\u002Fopenai-compatible\"\u003C\u002Fcode> and an \u003Ccode>options.baseURL\u003C\u002Fcode> such as \u003Ccode>http:\u002F\u002F127.0.0.1:8080\u002Fv1\u003C\u002Fcode> for \u003Ccode>llama-server\u003C\u002Fcode>.\u003C\u002Fp>\u003Cp>Built-in support for custom endpoints in Orca itself is still a request, not a feature. GitHub issues \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca\u002Fissues\u002F9335?ref=ghost.jdneer.com\">#9335\u003C\u002Fa> (OpenAI-compatible providers) and \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca\u002Fissues\u002F9239?ref=ghost.jdneer.com\">#9239\u003C\u002Fa> (Anthropic-compatible accounts) were both open as of October 18.\u003C\u002Fp>\u003Cp>Local models also bring the hardware question back. Mini PCs built on AMD's Ryzen AI Max+ 395 (Strix Halo) put up to 128GB of unified memory next to the GPU, enough to load a model like OpenAI's gpt-oss-120b, which \u003Ca href=\"https:\u002F\u002Fwww.servethehome.com\u002Fframework-desktop-review-a-solid-amd-strix-halo\u002F4\u002F?ref=ghost.jdneer.com\">ServeTheHome\u003C\u002Fa> ran at about 45 tokens per second in LM Studio on a Framework Desktop. Framework's DIY Edition configurator listed the 128GB Max+ 395 configuration at $3,449 on October 18 (out of stock at the time), a fraction of what a multi-GPU workstation with that much video memory costs. Memory bandwidth is the real limit: a \u003Ca href=\"https:\u002F\u002Fforum.level1techs.com\u002Ft\u002Fstrix-halo-ryzen-ai-max-395-llm-benchmark-results\u002F233796?ref=ghost.jdneer.com\">Level1Techs forum tester\u003C\u002Fa> measured about 215 GB\u002Fs out of a 256 GB\u002Fs theoretical peak, so large dense models generate slowly, and mixture-of-experts models are the better fit.\u003C\u002Fp>\u003Cp>One caveat applies whether the model is local or in the cloud: Orca turns agent permission prompts off by default. The \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fdocs\u002Fagents\u002Fsupported?ref=ghost.jdneer.com\">docs\u003C\u002Fa> say \"Orca pre-fills each supported CLI's permission-bypass flag for new launches,\" such as \u003Ccode>--dangerously-skip-permissions\u003C\u002Fcode> for Claude Code and \u003Ccode>--dangerously-bypass-approvals-and-sandbox\u003C\u002Fcode> for Codex. Anyone who wants agents to ask before running commands can switch to Manual under Settings, Agents, Agent Permissions.\u003C\u002Fp>\u003Cfigure class=\"kg-card kg-image-card kg-card-hascaption\">\u003Cimg src=\"\u002Fcontent\u002Fimages\u002F2026\u002F10\u002Forca-docs-permissions-default.png\" class=\"kg-image\" alt=\"Orca docs section titled Permissions default, listing the permission-bypass flags Orca pre-fills for each agent\" loading=\"lazy\" width=\"1608\" height=\"816\" srcset=\"\u002Fcontent\u002Fimages\u002Fsize\u002Fw600\u002F2026\u002F10\u002Forca-docs-permissions-default.png 600w, \u002Fcontent\u002Fimages\u002Fsize\u002Fw1000\u002F2026\u002F10\u002Forca-docs-permissions-default.png 1000w, \u002Fcontent\u002Fimages\u002Fsize\u002Fw1600\u002F2026\u002F10\u002Forca-docs-permissions-default.png 1600w, \u002Fcontent\u002Fimages\u002F2026\u002F10\u002Forca-docs-permissions-default.png 1608w\" sizes=\"(min-width: 720px) 720px\" \u002F>\u003Cfigcaption>The \"Permissions default\" section of Orca's supported agents page, captured October 18. Source: \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fdocs\u002Fagents\u002Fsupported?ref=ghost.jdneer.com\">Orca docs\u003C\u002Fa>.\u003C\u002Ffigcaption>\u003C\u002Ffigure>\u003Ch2 id=\"practical-takeaways\">Practical takeaways\u003C\u002Fh2>\u003Cp>\u003Cstrong>Who it suits:\u003C\u002Fstrong> developers who already pay for one or more coding agents, read every diff before merging, and regularly have several independent tasks going at once. It also fits people who want to check on agents from a phone.\u003C\u002Fp>\u003Cp>\u003Cstrong>Who should wait:\u003C\u002Fstrong> anyone who mostly makes small, linear changes with a single agent, needs VS Code extensions or a full debugger in the same window, or wants a tool that changes slowly. Teams rolling it out widely may want to pin versions given the release pace.\u003C\u002Fp>\u003Cp>\u003Cstrong>Caveats to plan for:\u003C\u002Fstrong> agents launch with permission prompts off unless Manual is selected, token use grows with every parallel agent, each worktree uses its own disk space, and review time becomes the limit. Packaged builds send anonymous usage telemetry, which the \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fdocs\u002Ftelemetry?ref=ghost.jdneer.com\">privacy docs\u003C\u002Fa> say never includes code, prompts or repo names, and which turns off when \u003Ccode>DO_NOT_TRACK=1\u003C\u002Fcode> or \u003Ccode>ORCA_TELEMETRY_DISABLED=1\u003C\u002Fcode> is set. The \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fenterprise?ref=ghost.jdneer.com\">enterprise page\u003C\u002Fa> lists SOC 2 \"readiness,\" not a completed audit.\u003C\u002Fp>\u003Cp>\u003Cstrong>How to try it:\u003C\u002Fstrong> download a build from the \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002F?ref=ghost.jdneer.com\">Orca site\u003C\u002Fa> or GitHub releases, or use a package manager. On macOS:\u003C\u002Fp>\u003Cpre>\u003Ccode>brew install --cask stablyai\u002Forca\u002Forca\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>On Arch Linux, via the AUR:\u003C\u002Fp>\u003Cpre>\u003Ccode>yay -S stably-orca-bin\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>On first launch, the \u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fdocs\u002Finstall?ref=ghost.jdneer.com\">install guide\u003C\u002Fa> says Orca offers to import existing \u003Ccode>~\u002F.claude\u003C\u002Fcode>, \u003Ccode>~\u002F.codex\u003C\u002Fcode> and Ghostty settings. On Linux the command line tool is named \u003Ccode>orca-ide\u003C\u002Fcode> to avoid clashing with the GNOME Orca screen reader. Lima's advice for a first run is a good one: start with two parallel sessions on a project you know well and see whether it clicks.\u003C\u002Fp>\u003Cp>\u003Cem>Sources: \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002F?ref=ghost.jdneer.com\">\u003Cem>Orca website\u003C\u002Fem>\u003C\u002Fa>\u003Cem>, \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fwww.onorca.dev\u002Fdocs?ref=ghost.jdneer.com\">\u003Cem>Orca docs\u003C\u002Fem>\u003C\u002Fa>\u003Cem>, \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca?ref=ghost.jdneer.com\">\u003Cem>stablyai\u002Forca on GitHub\u003C\u002Fem>\u003C\u002Fa>\u003Cem>, \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fangelo-lima.fr\u002Fen\u002Forca-ide-agentic-development-en\u002F?ref=ghost.jdneer.com\">\u003Cem>Angélo Lima\u003C\u002Fem>\u003C\u002Fa>\u003Cem>, \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fjames-carr.org\u002Fposts\u002F2026-09-30-current-agentic-development-setup-orca-oh-my-pi\u002F?ref=ghost.jdneer.com\">\u003Cem>James Carr\u003C\u002Fem>\u003C\u002Fa>\u003Cem>, \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fwww.ramonchancay.me\u002Fblog\u002Forca-parallel-coding-agent-fleet?ref=ghost.jdneer.com\">\u003Cem>Ramón Chancay\u003C\u002Fem>\u003C\u002Fa>\u003Cem>, \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fmartintapia.com\u002Fblog\u002F2026\u002Fi-tried-orca-and-vscode-agents-window-i-went-back-to-tmux-shortly-after?ref=ghost.jdneer.com\">\u003Cem>Martin Tapia\u003C\u002Fem>\u003C\u002Fa>\u003Cem>, \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fstablyai\u002Forca\u002Fissues\u002F23835?ref=ghost.jdneer.com\">\u003Cem>GitHub issue #23835\u003C\u002Fem>\u003C\u002Fa>\u003Cem>, \u003C\u002Fem>\u003Ca href=\"https:\u002F\u002Fnews.ycombinator.com\u002Fitem?id=47549197&ref=ghost.jdneer.com\">\u003Cem>Hacker News\u003C\u002Fem>\u003C\u002Fa>\u003Cem>.\u003C\u002Fem>\u003C\u002Fp>","Orca is a free, MIT-licensed desktop app from Stably AI that runs Claude Code, Codex and other coding agents side by side, each in its own git worktree....",{"src":60,"height":67,"width":68,"type":69},1335,2000,"image\u002Fjpeg",{"name":71,"url":72,"owner":13},"Aiko Tanaka",null,[],{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":75},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M5 12h14M5 12l6 6m-6-6l6-6\"\u002F>",{"left":15,"top":15,"width":26,"height":26,"rotate":15,"vFlip":13,"hFlip":13,"body":77},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M6.5 7.5a1 1 0 1 0 2 0a1 1 0 1 0-2 0\"\u002F>\u003Cpath d=\"M3 6v5.172a2 2 0 0 0 .586 1.414l7.71 7.71a2.41 2.41 0 0 0 3.408 0l5.592-5.592a2.41 2.41 0 0 0 0-3.408l-7.71-7.71A2 2 0 0 0 11.172 3H6a3 3 0 0 0-3 3\"\u002F>\u003C\u002Fg>"]