[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:whip":3},"\u003Ch1>\n  \u003Cbr \u002F>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fcontext-labs\u002Fwhip\u002FHEAD\u002Fapps\u002Fdesktop\u002Fresources\u002FWhip.png\" alt=\"WhipCode logo\" width=\"150\" \u002F>\u003C\u002Fa>\n   \u003Cbr \u002F>\n  WhipCode\n  \u003Cbr \u002F>\n\u003C\u002Fh1>\u003Cp align=\"center\">\u003Cstrong>A coding agent built for work that doesn’t fit in one context window.\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\u002Fblob\u002FHEAD\u002FLICENSE\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache_2.0-blue.svg\" alt=\"Apache 2.0 license\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\u002Freleases\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDesktop-beta-blue\" alt=\"Desktop beta\" \u002F>\u003C\u002Fa>\n\u003C\u002Fp>\u003Cp align=\"center\">\n  \u003Ca href=\"#quickstart\" rel=\"nofollow ugc noopener\">Quickstart\u003C\u002Fa> ·\n  \u003Ca href=\"#what-is-this\" rel=\"nofollow ugc noopener\">What is this?\u003C\u002Fa> ·\n  \u003Ca href=\"#how-it-works\" rel=\"nofollow ugc noopener\">How it works\u003C\u002Fa> ·\n  \u003Ca href=\"#benchmarks\" rel=\"nofollow ugc noopener\">Benchmarks\u003C\u002Fa> ·\n  \u003Ca href=\"#development\" rel=\"nofollow ugc noopener\">Development\u003C\u002Fa>\n\u003C\u002Fp>\u003Cp>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\u002Fblob\u002FHEAD\u002Fdocs\u002Fassets\u002Fwhipcode-desktop.png\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fcontext-labs\u002Fwhip\u002FHEAD\u002Fdocs\u002Fassets\u002Fwhipcode-desktop.png\" alt=\"WhipCode Desktop with a conversation, three working subagents, and an execution trace timeline side by side\" \u002F>\u003C\u002Fa>\u003C\u002Fp>\n\u003Ch2>Quickstart\u003C\u002Fh2>\n\u003Cp>\u003Cstrong>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\u002Freleases\" rel=\"nofollow ugc noopener\">Download Desktop Beta\u003C\u002Fa>\u003C\u002Fstrong> — Apple Silicon · macOS 14 or newer.\u003C\u002Fp>\n\u003Cp>Use a release built after the clean-project reset; old Desktop releases are not an upgrade path.\u003C\u002Fp>\n\u003Col>\n\u003Cli>Download \u003Ccode>whipcode-desktop-darwin-arm64.dmg\u003C\u002Fcode>, open it, and drag \u003Cstrong>Whip Beta\u003C\u002Fstrong> into Applications.\u003C\u002Fli>\n\u003Cli>Launch it and choose \u003Cstrong>Set up this Mac\u003C\u002Fstrong> if prompted. The app includes its matching backend; no separate CLI installation is needed.\u003C\u002Fli>\n\u003Cli>Connect a model provider, open your project, and describe what you want to do.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Cp>Bring an API key or use a supported subscription login. Provider access is separate from installing WhipCode. See \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\u002Fblob\u002FHEAD\u002Fdocs\u002Fmodels-providers.md\" rel=\"nofollow ugc noopener\">providers and authentication\u003C\u002Fa>.\u003C\u002Fp>\n\u003Cp>Prefer the terminal? Install the standalone CLI on macOS or Linux:\u003C\u002Fp>\n\u003Cpre>\u003Ccode class=\"language-sh\">curl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fcontext-labs\u002Fwhip\u002Fmain\u002Finstall.sh | sh\nwhipcode\n\u003C\u002Fcode>\u003C\u002Fpre>\n\u003Cp>For an exact version, use the release's \u003Ccode>install.sh\u003C\u002Fcode> command from its notes—no\nversion environment variable is needed for releases with the two-script split.\nThe companion \u003Ccode>latest.sh\u003C\u002Fcode> always selects stable v1+, never an alpha.\u003C\u002Fp>\n\u003Cp>The new CLI track starts at \u003Ccode>v1.0.0-alpha.N\u003C\u002Fcode>. Prerelease installation is explicit;\ndefault stable installation becomes available with \u003Ccode>v1.0.0\u003C\u002Fcode>. Existing internal\ninstalls must follow the \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\u002Fblob\u002FHEAD\u002Fdocs\u002Fteam-reset.md\" rel=\"nofollow ugc noopener\">manual reset checklist\u003C\u002Fa>, not upgrade in place.\u003C\u002Fp>\n\u003Cp>Desktop-managed installations update through Desktop; standalone CLI installations use \u003Ccode>whipcode update\u003C\u002Fcode>. \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\u002Fblob\u002FHEAD\u002Fdocs\u002Fsetup.md#desktop-installation-and-upgrades\" rel=\"nofollow ugc noopener\">Setup and upgrades →\u003C\u002Fa>\u003C\u002Fp>\n\u003Cp>For the web client, start the daemon with \u003Ccode>whipcode daemon start\u003C\u002Fcode>, then run\n\u003Ccode>whipcode web\u003C\u002Fcode>. It opens your browser and stays running as a separate gateway;\nCtrl+C stops web access, not daemon work. Ordinary daemon and Desktop startup\nopen no web listener. See \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcontext-labs\u002Fwhip\u002Fblob\u002FHEAD\u002Fdocs\u002Fweb-app.md#run-the-packaged-application-locally\" rel=\"nofollow ugc noopener\">web access\u003C\u002Fa>\nfor \u003Ccode>--no-open\u003C\u002Fcode>, managed startup, and trusted-network configuration.\u003C\u002Fp>\n\u003Ch2>What is this?\u003C\u002Fh2>\n\u003Cp>WhipCode is an open-source coding agent with a Desktop app, terminal UI, and web client. A shared daemon owns the work, so closing a window doesn’t end the session.\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Delegate recursively.\u003C\u002Fstrong> Agents split work into focused tasks, launch their own agents, and coordinate through messages.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Keep context within reach.\u003C\u002Fstrong> Search and read history, files, and large tool outputs in bounded slices instead of cramming everything into a prompt.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>See what’s happening.\u003C\u002Fstrong> Follow conversations, inspect the execution tree and timeline, and track model usage and cost.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Stay in control.\u003C\u002Fstrong> Choose models, set tool permissions, and connect local or remote execution hosts. Extend agents with skills and MCP servers.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>How it works\u003C\u002Fh2>\n\u003Col>\n\u003Cli>\u003Cstrong>Describe the task.\u003C\u002Fstrong> Give an agent a goal and the context it needs.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Let it work.\u003C\u002Fstrong> A recursive language model (RLM) loop uses short Starlark or JavaScript programs to inspect context, call tools, and delegate—not just a growing chat transcript.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Inspect and steer.\u003C\u002Fstrong> Follow the trace\u003C\u002Fli>\n\u003C\u002Fol>\n",1790633833833]