+
-
+
supervise multiple coding agents in one terminal.
- -herd your agents.
-- herdr.dev · install · quick start · supported agents · integrations · configuration · agent skill · socket api + herdr.dev · install · quick start · supported agents · integrations · configuration · socket api
--- -herdr is a terminal-native workspace manager for coding agents. - -run Claude Code, Codex, pi, opencode, droid, amp and plain shells side by side in your existing terminal. herdr gives you workspaces, tabs, panes, automatic agent detection, and notification alerts so you can see which agent is blocked, done, or still working without leaving the command line. - -it runs inside ghostty, alacritty, kitty, wezterm, and even inside tmux. it is a single Rust binary, not a separate GUI window, electron wrapper, or web dashboard. - https://github.com/user-attachments/assets/043ec09f-4bdd-41d5-aee0-8fda6b83e267 -## why herdr +**agent multiplexer that lives in your terminal.** -running one coding agent is easy. running several in parallel gets messy fast. +workspaces, tabs, panes. mouse-native: click, drag, split. every agent at a glance: blocked, working, done. detach and reattach, agents keep running. no gui app, no electron, no mac-only native wrapper. you see the agent's own terminal, not someone's interpretation of it. -most tools in this space either replace your environment or give you panes without any awareness of what those panes are doing. herdr stays inside your terminal and adds the missing layer: supervision. - -with herdr you can: - -- run multiple coding agents in parallel in one terminal-native workspace -- scan workspaces quickly and see which one needs attention -- spot whether an agent is blocked, finished, working, or idle -- jump between repos, tabs, and panes without losing context -- let agents create panes, spawn helpers, read output, and wait on each other through the local api +--- ## install @@ -42,15 +25,11 @@ with herdr you can: curl -fsSL https://herdr.dev/install.sh | sh ``` -or download the binary directly from [releases](https://github.com/ogulcancelik/herdr/releases). - -requirements: linux or macos. +or download the binary from [releases](https://github.com/ogulcancelik/herdr/releases). requires linux or macos. ### update -herdr checks for updates automatically in the background. when a new version is ready, you'll see a notification in the ui. restart to apply it. - -for a manual update: +herdr notifies you when a new version is available. run manually to update: ```bash herdr update @@ -58,82 +37,102 @@ herdr update ## quick start -launch herdr: - ```bash herdr ``` -then do this: +by default herdr launches or attaches to a background session server. `ctrl+b q` detaches the client. agents keep running. use `herdr server stop` to stop the server. use `--no-session` for the old single-process mode. 1. press `n` to create a workspace 2. run an agent in the root pane 3. press `ctrl+b` to enter navigate mode 4. use `v` or `-` to split panes, or `c` to create a new tab -5. run more agents side by side -6. watch the sidebar to see which workspace or agent needs attention +5. watch the sidebar for blocked, working, and done states -on first run, herdr opens a short onboarding flow so you can choose your notification style. after that, if a session is restored you'll land in terminal mode; otherwise you'll start in **navigate mode**. +on first run herdr opens a short onboarding flow. after that, restored sessions land in terminal mode; fresh sessions start in **navigate mode**. + +## how it compares + +| | tmux | gui managers | herdr | +|--------------------------|------|--------------|-------| +| persistent sessions | ✓ | — | ✓ | +| detach / reattach | ✓ | — | ✓ | +| panes, tabs, workspaces | ✓ | ✓ | ✓ | +| agent awareness | — | ✓ | ✓ | +| lives in your terminal | ✓ | — | ✓ | +| real terminal views | ✓ | — | ✓ | +| mouse-native | — | ✓ | ✓ | +| lightweight binary | ✓ | — | ✓ | +| agents can orchestrate | ? | ? | ✓ | + +tmux gives you persistence and panes, but it was built before agents existed. gui managers show agent state, but they make you leave your terminal and use their wrapped view. herdr is persistence and awareness in one tool that stays out of your way. + +## persistence + +start herdr on your desktop or server. run your agents, split panes, do your work. press `ctrl+b q` to detach. close your terminal, close your laptop; your agents keep running. open a new terminal, run `herdr`, you're back. same session, same panes, same agents. + +### from anywhere + +need to check on your agents from your phone? just ssh in and run herdr. any ssh client works. no app to download, no account to create. + +``` +ssh you@yourserver +herdr +``` + +same session, same agents, same state. + +## agent awareness + +the sidebar shows which agents are blocked, working, or done. workspaces roll up to their most urgent state so you can scan the full list at a glance. + +states: + +- 🔴 **blocked** — agent needs input or approval +- 🟡 **working** — agent is actively running +- 🔵 **done** — work finished, you have not looked at it yet +- 🟢 **idle** — done and seen + +detection works by reading foreground process and terminal output. zero config, no hooks required. for agents that expose hooks, the socket api integration gives more robust state reporting. + +## lives in your terminal + +not a gui window, not a web dashboard, not electron. herdr runs inside whatever terminal you already use. single rust binary, no dependencies. works inside tmux. ## what you get -### workspaces, tabs, and panes +- **workspaces** — organized around git repos or folder names, each with its own tabs and panes +- **tabs** — first-class in the socket api and cli +- **mouse-native** — click panes, drag borders, select text to copy; not keyboard-only +- **notifications** — sounds and toasts for background events; tab-aware suppression +- **9 built-in themes** — catppuccin (default), tokyo night, dracula, nord, gruvbox, one dark, solarized, kanagawa, rosé pine +- **session persistence** — pane processes survive client detach; sessions restore after full restart -herdr is organized around workspaces. each workspace can contain multiple tabs, and each tab can contain multiple panes. +## agents can use herdr too -workspaces open immediately as real terminal contexts. the first pane in a workspace is the **root pane**. it anchors the workspace and gives it its default identity: +the local unix socket lets agents create workspaces, split panes, spawn helpers, read output, and wait for state changes. -- if the root pane is inside a git repo, the workspace label defaults to the repo name -- otherwise it falls back to the current folder name -- manual rename is supported, but the default model is repo or folder first +```bash +# create a workspace and tab +herdr workspace create --cwd ~/project --label "api" +herdr tab create --label "logs" -this keeps workspaces lightweight. they are project contexts first, labels second. +# split a pane and run +herdr pane split 1-1 --direction right +herdr pane run 1-2 "npm test" -### awareness and notifications +# wait for an agent +herdr wait agent-status 1-1 --status done -the sidebar is split into two layers: +# read output +herdr pane read 1-2 --source recent --lines 50 +``` -- **top:** workspaces, each with one aggregate state dot -- **bottom:** detected agents inside the selected workspace, or across all workspaces when you switch the agent panel scope - -herdr automatically detects running agents by looking at the foreground process and reading terminal output. workspace rollups surface the most urgent thing happening in each workspace so you can scan the full list quickly. - -in expanded view, the workspace list and agent list have their own resizable sections. in collapsed view, you still get compact per-pane agent indicators. - -states map to: - -- 🔴 **blocked** — agent needs input or approval -- 🔵 **done** — work finished and you have not looked at it yet -- 🟡 **working** — agent is actively running -- 🟢 **idle** — done, seen, and calm - -rollups prefer the most urgent state in a workspace: blocked first, then unseen finished work, then working, then idle. - -plain shells still matter to the workspace itself, but the sidebar stays focused on actual agents. - -if ambient sidebar awareness is not enough, herdr can also play sounds or show top-right toast notifications for background events. notification suppression is tab-aware: the active tab stays quiet, but background tabs in the same workspace can still alert. - -### agents can use herdr too - -herdr is not only for humans supervising agents. it is also becoming a shared control surface for the agents themselves. - -agents running inside herdr can use the local socket api and cli wrappers to: - -- create new workspaces for parallel tasks -- create, focus, rename, and close tabs -- split panes for servers, logs, tests, or scratch work -- spawn other agents in sibling panes -- read pane output or wait for output matches -- send text and keys into other panes -- wait for another agent to finish before continuing - -for that workflow, start with [`SKILL.md`](./SKILL.md) if you want a reusable agent skill, or [`SOCKET_API.md`](./SOCKET_API.md) if you want the low-level socket protocol and cli wrapper reference. +full reference: [`SOCKET_API.md`](./SOCKET_API.md) and [`SKILL.md`](./SKILL.md). ## supported agents -herdr detects supported agents automatically with zero setup. it identifies the foreground process in each pane and reads the live bottom of the terminal buffer to infer agent state. for agents that expose hooks or plugins, direct integrations are the more robust path because they forward semantic state to herdr over the local socket api. - -the following agents have been tested: +automatic detection works out of the box. process name matching plus terminal output heuristics. | agent | idle / done | working | blocked | |-------|-------------|---------|---------| @@ -144,30 +143,13 @@ the following agents have been tested: | [amp](https://ampcode.com) | ✓ | ✓ | ✓ | | [opencode](https://github.com/anomalyco/opencode) | ✓ | ✓ | ✓ | -heuristics also exist for these agents but have not been fully tested yet: +detected but not fully tested: gemini cli, cursor agent, cline, kimi, github copilot cli. -- [gemini cli](https://github.com/google-gemini/gemini-cli) -- [cursor agent](https://cursor.com/cli) -- [cline](https://github.com/cline/cline) -- [kimi](https://kimi.ai) -- [github copilot cli](https://cli.github.com) +for agents outside the built-in list, herdr still works as a terminal multiplexer with workspaces, panes, and tiling. custom integrations can report agent labels over the socket api. see [`SOCKET_API.md`](./SOCKET_API.md). -for any other cli agent, herdr still works as a terminal-native multiplexer. you still get workspaces, panes, tiling, and notifications even when richer state detection is not available yet. +### direct integrations -for agents outside the built-in list, herdr also supports custom hook or plugin integrations over the local socket api. a custom integration can report any non-empty agent label, and that label will appear in pane info, events, and the sidebar without needing first-party built-in detection. see [`SOCKET_API.md`](./SOCKET_API.md) for the raw api surface. - -## optional direct integrations - -automatic detection works out of the box. if an agent exposes hooks or plugins, the better path is to let it report state to herdr over the local socket api. - -herdr ships one-command integrations for: - -- [pi](./INTEGRATIONS.md#pi) -- [claude code](./INTEGRATIONS.md#claude-code) -- [codex](./INTEGRATIONS.md#codex) -- [opencode](./INTEGRATIONS.md#opencode) - -install them with: +the built-in pi, claude code, codex, and opencode integrations forward semantic state to herdr over the socket api. install with: ```bash herdr integration install pi @@ -176,170 +158,45 @@ herdr integration install codex herdr integration install opencode ``` -remove them with: +see [`INTEGRATIONS.md`](./INTEGRATIONS.md) for setup details. -```bash -herdr integration uninstall pi -herdr integration uninstall claude -herdr integration uninstall codex -herdr integration uninstall opencode -``` +## keybindings -these integrations forward semantic state to herdr over the local socket api. that gives you more robust status reporting, but it does not replace herdr's core process detection model. for setup details, installed file locations, caveats, and the exact bundled hook or plugin source files, see [`INTEGRATIONS.md`](./INTEGRATIONS.md). - -known codex caveat: codex currently renders hook lifecycle lines in its own tui when hooks are enabled. that noise is upstream codex behavior, not herdr-specific. - -## usage - -### navigate mode - -navigate mode is the workspace control layer. press `ctrl+b` to enter it. - -movement actions stay in navigate mode. mutating actions like split, close, new workspace, new tab, and sidebar toggle return you to terminal mode. - -common defaults: - -- `n` new workspace -- `shift+n` rename workspace -- `d` close workspace -- `c` new tab -- `v` / `-` split pane -- `x` close pane -- `f` fullscreen -- `r` resize mode -- `b` toggle sidebar - -optional direct bindings are available but **unset by default**. you can bind workspace, tab, and pane switching directly in terminal mode without going through the prefix first. - -prefix mode also supports custom commands with `[[keys.command]]`. that lets you launch detached shell helpers or temporary overlay panes. two common examples are opening `lazygit` in a pane and running your own script with the active herdr ids and cwd. - -example: - -```toml -[keys] -previous_workspace = "ctrl+alt+[" -next_workspace = "ctrl+alt+]" -previous_tab = "alt+[" -next_tab = "alt+]" -focus_pane_left = "alt+h" -focus_pane_down = "alt+j" -focus_pane_up = "alt+k" -focus_pane_right = "alt+l" - -[[keys.command]] -key = "g" -type = "pane" -command = "lazygit" - -[[keys.command]] -key = "o" -command = "~/bin/herdr-open-current \"$HERDR_ACTIVE_WORKSPACE_ID\" \"$HERDR_ACTIVE_TAB_ID\" \"$HERDR_ACTIVE_PANE_ID\" \"$HERDR_ACTIVE_PANE_CWD\"" -``` - -full keybinding and config reference: [`CONFIGURATION.md`](./CONFIGURATION.md) - -### resize mode +press `ctrl+b` to enter navigate mode. | key | action | |-----|--------| -| `h` `l` | resize width | -| `j` `k` | resize height | -| `esc` | exit resize mode | +| `n` | new workspace | +| `shift+n` | rename workspace | +| `shift+d` | close workspace | +| `c` | new tab | +| `v` / `-` | split pane | +| `x` | close pane | +| `b` | toggle sidebar | +| `f` | fullscreen pane | +| `r` | resize mode | +| `q` | detach (quit client) | -### mouse +resize mode: `h`/`l` resize width, `j`/`k` resize height, `esc` exit. -mouse support is built in. herdr is not keyboard-only. - -- click a workspace in the sidebar to switch -- click tabs to switch within the active workspace -- click the tab-bar `<` / `>` buttons to scroll overflowing tabs -- drag tabs to reorder them -- click a pane to focus it -- drag split borders to resize -- drag in a pane to select text; release to copy it to your system clipboard -- right-click a workspace for a context menu -- scroll in the sidebar to navigate workspaces -- click `«` / `»` at the sidebar bottom to collapse or expand it - -text copy uses OSC 52, so it depends on your terminal's clipboard support. - -### terminal mode - -terminal mode is a real terminal. your shell, vim, htop, ssh, and other full-screen tools work normally. press the prefix key (`ctrl+b`) to go back to navigate mode. - -## docs map - -use the dedicated docs for detailed setup and automation work: - -- [`CONFIGURATION.md`](./CONFIGURATION.md) — config file, keybindings, themes, notifications, onboarding, ui options, and environment variables -- [`INTEGRATIONS.md`](./INTEGRATIONS.md) — install and behavior notes for pi, Claude Code, Codex, and opencode integrations -- [`SKILL.md`](./SKILL.md) — reusable agent skill for agents already running inside herdr -- [`SOCKET_API.md`](./SOCKET_API.md) — canonical socket protocol and the `herdr workspace`, `herdr tab`, `herdr pane`, and `herdr wait` cli wrappers +mouse is supported throughout. full reference: [`CONFIGURATION.md`](./CONFIGURATION.md). ## configuration config file: `~/.config/herdr/config.toml` -print the full default config with: - ```bash -herdr --default-config +herdr --default-config # print full default config ``` -### themes +in-app settings screen for theme, sound, and toast preferences. full reference: [`CONFIGURATION.md`](./CONFIGURATION.md). -herdr ships with 9 built-in themes: catppuccin (default), tokyo night, dracula, nord, gruvbox, one dark, solarized, kanagawa, and rosé pine. +## docs -```toml -[theme] -name = "tokyo-night" -``` - -you can also override individual color tokens on top of any base theme. see [`CONFIGURATION.md`](./CONFIGURATION.md) for the full token reference. - -for day-to-day changes, you do not have to edit the config by hand. herdr also has an in-app settings screen for theme, sound, and toast preferences. - -## session persistence - -herdr saves your workspaces, tabs, pane layouts, pane working directories, and focused tab or pane automatically as you work, then restores them on restart. sessions are stored at `~/.config/herdr/session.json`. - -through the socket api, tabs are first-class too. raw integrations and cli wrappers can list, create, focus, rename, and close tabs directly while pane ids stay workspace-scoped. - -use `--no-session` to start fresh. - -## how agent detection works - -herdr does not require hooks or agent-side configuration for built-in detection. automatic detection works by: - -1. identifying the foreground process of each pane's pty via `/proc` on linux or `proc_pidinfo` on macos -2. matching the process name against known agents -3. reading the live bottom of the terminal buffer and applying per-agent heuristics to determine state - -this means detection works with any supported agent, installed any way, with zero setup. if it runs in a terminal, herdr can see it. - -when an agent exposes hooks or plugins, the more robust option is to forward state to herdr over the local socket api. that is what the built-in pi, claude code, codex, and opencode integrations do. - -the heuristics are matched against each agent's real terminal output: prompt boxes, spinners, waiting-for-input messages, and tool execution indicators. detection runs on a separate async task per pane, polled every 300 to 500 ms, decoupled from terminal rendering. - -## cli wrappers - -herdr's workspace, tab, pane, and wait commands are documented in [`SOCKET_API.md`](./SOCKET_API.md) together with the socket methods they wrap. - -both `workspace create` and `tab create` support optional `--label` flags, so scripts and agents can name contexts immediately instead of renaming them after creation. both create commands also return the created root pane in their json response, so clients can act on the new pane without an extra lookup. - -workspace ids are compact public ids like `1`, `2`, `3`. -tab ids are compact public ids like `1:1`, `1:2`, `2:1`. -pane ids are compact public ids like `1-1`, `1-2`, `2-1`. - -even with tabs enabled, pane ids remain workspace-scoped public ids rather than `workspace-tab-pane` triples. both tab ids and pane ids are positional within the current live session, so numbering compacts when tabs, workspaces, or panes are closed. - -## built with agents - -i had never written Rust before starting this project. herdr was built almost entirely through ai coding agents, the same ones it is designed to multiplex. i supervised the architecture and specs. agents wrote the code. - -that is part of the point here. herdr is a useful tool, but it is also a proof of what current coding agents can build when they are directed carefully inside a real workflow. - -there will be rough edges. if you hit one, [open an issue](https://github.com/ogulcancelik/herdr/issues). +- [`CONFIGURATION.md`](./CONFIGURATION.md) — keybindings, themes, notifications, environment variables +- [`INTEGRATIONS.md`](./INTEGRATIONS.md) — pi, claude code, codex, opencode integrations +- [`SKILL.md`](./SKILL.md) — reusable agent skill +- [`SOCKET_API.md`](./SOCKET_API.md) — socket protocol and cli reference ## building from source @@ -353,10 +210,10 @@ cargo build --release ## testing ```bash -just test # unit tests -just test-all # full local test suite +just test # unit tests +just test-all # full local test suite ``` ## license -AGPL-3.0: free to use, modify, and distribute. if you distribute a modified version, you must open-source your changes under the same license. +AGPL-3.0: free to use, modify, and distribute. modified versions must be open-sourced under the same license. diff --git a/website/assets/sidebar-agents.png b/website/assets/sidebar-agents.png index 1d5c5041..06e03fb2 100644 Binary files a/website/assets/sidebar-agents.png and b/website/assets/sidebar-agents.png differ diff --git a/website/assets/sidebar-workspaces.png b/website/assets/sidebar-workspaces.png index a8160f1e..2d5aa629 100644 Binary files a/website/assets/sidebar-workspaces.png and b/website/assets/sidebar-workspaces.png differ diff --git a/website/css/style.css b/website/css/style.css new file mode 100644 index 00000000..54061ec9 --- /dev/null +++ b/website/css/style.css @@ -0,0 +1,719 @@ +*, +*::before, +*::after { + margin: 0; + padding: 0; + box-sizing: border-box; + scrollbar-width: thin; + scrollbar-color: var(--border-bright) var(--mantle); +} + +*::-webkit-scrollbar { + width: 10px; + height: 10px; +} + +*::-webkit-scrollbar-track { + background: var(--mantle); +} + +*::-webkit-scrollbar-thumb { + background: var(--border-bright); + border: 2px solid var(--mantle); + border-radius: 999px; +} + +*::-webkit-scrollbar-thumb:hover { + background: var(--text-dim); +} + +*::-webkit-scrollbar-corner { + background: var(--mantle); +} + +:root { + --base: #11111b; + --surface: #1e1e2e; + --mantle: #181825; + --overlay: #262637; + --border: #313244; + --border-bright: #45475a; + --white: #cdd6f4; + --text: #a6adc8; + --text-dim: #6c7086; + --text-faint: #45475a; + --green: #a6e3a1; + --blue: #89b4fa; + --red: #f38ba8; + --yellow: #f9e2af; + --mauve: #cba6f7; + --max-w: 880px; +} + +html { + background: var(--base); + color: var(--text); + font-family: "JetBrains Mono", monospace; + font-size: 17px; + -webkit-font-smoothing: antialiased; + scroll-behavior: smooth; +} + +body { + overflow-x: clip; +} + +a { + color: var(--text); + text-decoration: none; + border-bottom: 1px solid var(--border); + transition: + color 0.15s, + border-color 0.15s; +} +a:hover { + color: var(--green); + border-color: var(--green); +} + +::selection { + background: var(--overlay); + color: var(--white); +} + +/* ── State dots ── */ +.dot { + display: inline-block; + width: 8px; + height: 8px; + border-radius: 50%; + flex-shrink: 0; +} +.dot-blocked { + background: var(--red); +} +.dot-working { + background: var(--yellow); +} +.dot-done { + background: var(--blue); +} +.dot-idle { + background: var(--green); +} + +.text-blocked { + color: var(--red); +} +.text-working { + color: var(--yellow); +} +.text-done { + color: var(--blue); +} +.text-idle { + color: var(--green); +} + +/* ── Sections ── */ +.section { + max-width: var(--max-w); + margin: 0 auto; + padding: 0 2rem; +} + +.section-rule { + border: none; + border-top: 1px solid var(--border); + max-width: var(--max-w); + margin: 0 auto; +} + +.gap { + height: 5rem; +} +.gap-lg { + height: 7rem; +} +.gap-sm { + height: 3rem; +} + +.section-heading { + display: flex; + align-items: center; + gap: 0.65rem; + margin-bottom: 2rem; +} + +.section-heading h2 { + font-size: 0.75rem; + font-weight: 600; + text-transform: uppercase; + letter-spacing: 0.15em; + color: var(--text-dim); +} + +/* ── Hero ── */ +.hero { + position: relative; + max-width: var(--max-w); + margin: 0 auto; + padding: 6rem 2rem 0; + overflow: hidden; +} + +.hero-pasture { + position: absolute; + inset: 0; + overflow: hidden; + pointer-events: none; + z-index: 0; +} + +.hero > *:not(.hero-pasture) { + position: relative; + z-index: 1; +} + +.hero .sheep-el { + position: absolute; + font-family: "JetBrains Mono", monospace; + line-height: 1.15; + white-space: pre; + pointer-events: none; + user-select: none; + will-change: transform; +} + +.hero-logo { + width: 56px; + height: 56px; + border-radius: 14px; + margin-right: 1rem; + vertical-align: middle; +} + +.hero-name { + font-size: 1.5rem; + font-weight: 500; + text-transform: lowercase; + letter-spacing: 0.05em; + color: var(--white); + margin-bottom: 2rem; + display: inline-block; +} + +.hero-tagline { + font-size: 2.75rem; + font-weight: 700; + color: var(--white); + letter-spacing: -0.03em; + line-height: 1.1; + margin-bottom: 1.5rem; +} + +.hero-dots { + display: flex; + gap: 0.6rem; + margin-bottom: 3rem; +} + +.hero-dots .dot { + width: 10px; + height: 10px; +} + +.manifesto { + font-size: 1rem; + line-height: 1.9; + color: var(--text-dim); + margin-bottom: 3rem; +} + +.manifesto p + p { + margin-top: 1rem; +} +.manifesto strong { + color: var(--white); + font-weight: 500; +} + +.install-row { + display: flex; + align-items: center; + gap: 0.75rem; + margin-bottom: 1rem; + flex-wrap: wrap; +} + +.install-box { + display: inline-flex; + align-items: center; + gap: 0.5rem; + background: var(--mantle); + border: 1px solid var(--border); + border-radius: 6px; + padding: 0.55rem 0.9rem; + font-size: 0.875rem; + max-width: 100%; +} + +.install-box .prompt { + color: var(--green); +} +.install-box code { + color: var(--text); + white-space: normal; + overflow-wrap: anywhere; + word-break: break-word; +} + +.copy-btn { + background: none; + border: 1px solid var(--border); + border-radius: 4px; + color: var(--text-dim); + font-family: inherit; + font-size: 0.75rem; + padding: 0.35rem 0.6rem; + cursor: pointer; + transition: + color 0.15s, + border-color 0.15s; +} +.copy-btn:hover { + border-color: var(--text-dim); + color: var(--text); +} +.copy-btn.copied { + border-color: var(--green); + color: var(--green); +} + +.hero-meta { + font-size: 0.7rem; + color: var(--text-faint); + overflow-wrap: anywhere; +} + +.hero-meta a { + color: var(--text-dim); + border: none; +} +.hero-meta a:hover { + color: var(--green); +} + +/* ── Demo ── */ +.demo { + max-width: var(--max-w); + margin: 0 auto; + padding: 0 2rem; +} + +.demo-frame { + border: 1px solid var(--border); + border-radius: 8px; + overflow: hidden; +} + +.demo-frame video { + display: block; + width: 100%; + height: auto; +} + +/* ── Comparison ── */ +.position-table { + width: 100%; + border-collapse: collapse; + font-size: 0.875rem; +} + +.position-table th { + text-align: left; + font-size: 0.75rem; + font-weight: 600; + text-transform: uppercase; + letter-spacing: 0.1em; + color: var(--text-faint); + padding: 0.6rem 1rem; + border-bottom: 1px solid var(--border); + background: var(--mantle); +} + +.position-table th.herdr-col { + color: var(--green); +} + +.position-table td { + padding: 0.5rem 1rem; + color: var(--text-dim); + line-height: 1.55; + border-bottom: 1px solid var(--border); + vertical-align: top; +} + +.position-table tr:last-child td { + border-bottom: none; +} +.position-table td:first-child { + width: 11rem; + color: var(--text); + font-weight: 500; +} +.position-table td.herdr-col { + color: var(--text); +} + +.mark { + color: var(--text-faint); +} +.mark-yes { + color: var(--green); +} + +/* ── Prose block ── */ +.prose-block { + font-size: 0.875rem; + line-height: 1.85; + color: var(--text-dim); +} + +.prose-block strong { + color: var(--white); + font-weight: 500; +} + +/* ── Detach flow ── */ +.detach-flow { + background: var(--mantle); + border: 1px solid var(--border); + border-radius: 6px; + padding: 1.75rem 2rem; +} + +.detach-line { + font-size: 0.875rem; + line-height: 2.2; + color: var(--text-dim); +} + +.detach-line .prompt { + color: var(--green); +} +.detach-line .cmd { + color: var(--white); +} +.detach-line .key { + color: var(--blue); +} +.detach-line .comment { + color: var(--text-faint); +} +.detach-line .dim { + color: var(--text-faint); +} + +/* ── Awareness ── */ +.awareness-split { + display: grid; + grid-template-columns: 1fr; + gap: 1.5rem; + margin-bottom: 1.5rem; +} + +.awareness-panel { + display: flex; + gap: 1.25rem; + align-items: flex-start; +} + +.awareness-panel-img { + flex-shrink: 0; + width: 240px; + border: 1px solid var(--border); + border-radius: 4px; + overflow: hidden; +} + +.awareness-panel-img img { + display: block; + width: 100%; + height: auto; + image-rendering: auto; +} + +.awareness-panel-text h3 { + font-size: 0.875rem; + font-weight: 500; + color: var(--white); + margin-bottom: 0.35rem; +} + +.awareness-panel-text p { + font-size: 0.875rem; + color: var(--text-dim); + line-height: 1.65; +} + +.state-bar { + display: flex; + gap: 2rem; + padding: 0.5rem 0 0; +} + +.state-bar-item { + display: flex; + align-items: center; + gap: 0.5rem; + font-size: 0.875rem; + color: var(--text-dim); +} + +/* ── Terminal ── */ +.terminal-block { + background: var(--mantle); + border: 1px solid var(--border); + border-radius: 6px; + padding: 1.75rem 2rem; +} + +.terminal-block p { + font-size: 0.875rem; + line-height: 1.85; + color: var(--text-dim); +} + +.terminal-block strong { + color: var(--white); + font-weight: 500; +} + +.term-pills { + display: flex; + gap: 0.4rem; + flex-wrap: wrap; + margin-top: 1.25rem; +} + +.term-pill { + font-size: 0.75rem; + color: var(--text-dim); + background: var(--base); + padding: 0.2rem 0.6rem; + border: 1px solid var(--border); + border-radius: 4px; +} + +/* ── Comparison ── */ +.table-wrap { + width: 100%; + border: 1px solid var(--border); + border-radius: 6px; + overflow-x: auto; + overflow-y: hidden; + -webkit-overflow-scrolling: touch; +} + +.agents-table { + width: 100%; + border-collapse: collapse; + font-size: 0.875rem; +} + +.agents-table th { + text-align: left; + padding: 0.6rem 1rem; + color: var(--text-faint); + font-weight: 600; + font-size: 0.75rem; + text-transform: uppercase; + letter-spacing: 0.1em; + background: var(--mantle); + border-bottom: 1px solid var(--border); +} + +.agents-table td { + padding: 0.5rem 1rem; + border-bottom: 1px solid var(--border); + color: var(--text-dim); +} + +.agents-table tr:last-child td { + border-bottom: none; +} +.agents-table td:first-child { + color: var(--text); + font-weight: 500; +} + +.agents-table a { + border-bottom: 1px solid var(--border-bright); +} +.agents-table a:hover { + border-color: var(--green); +} + +.check { + color: var(--green); +} +.partial { + color: var(--yellow); +} +.untested { + color: var(--text-faint); + font-size: 0.75rem; +} + +/* ── API ── */ +.api-grid { + display: grid; + grid-template-columns: minmax(0, 0.82fr) minmax(0, 1.18fr); + gap: 2rem; + align-items: start; +} + +.api-grid > * { + min-width: 0; +} + +.api-text p { + font-size: 0.875rem; + color: var(--text-dim); + line-height: 1.85; + margin-bottom: 0.75rem; +} + +.api-text strong { + color: var(--white); + font-weight: 500; +} + +.api-links { + display: flex; + gap: 0.5rem; + margin-top: 0.75rem; + flex-wrap: wrap; +} + +.api-links a { + font-size: 0.75rem; + color: var(--text-dim); + padding: 0.3rem 0.65rem; + border: 1px solid var(--border); + border-radius: 4px; + transition: + color 0.15s, + border-color 0.15s; +} +.api-links a:hover { + border-color: var(--green); + color: var(--green); +} + +.code-block { + background: var(--mantle); + border: 1px solid var(--border); + border-radius: 6px; + padding: 1.25rem 1.5rem; + overflow-x: auto; + -webkit-overflow-scrolling: touch; +} + +.code-block pre { + font-size: 0.72rem; + line-height: 1.9; + color: var(--text-dim); +} + +.code-block .comment { + color: var(--text-faint); +} +.code-block .cmd { + color: var(--green); +} +.code-block .flag { + color: var(--blue); +} +.code-block .str { + color: var(--yellow); +} + +/* ── Footer ── */ +.footer { + max-width: var(--max-w); + margin: 0 auto; + padding: 2rem 2rem 4rem; + text-align: center; +} + +.footer p { + font-size: 0.7rem; + color: var(--text-faint); + line-height: 2; +} + +.footer a { + color: var(--text-dim); + border: none; +} +.footer a:hover { + color: var(--green); +} + +/* ── Responsive ── */ +@media (max-width: 768px) { + .hero, + .section, + .demo, + .footer { + padding-left: 1.25rem; + padding-right: 1.25rem; + } + .hero { + padding-top: 4rem; + } + .hero-tagline { + font-size: 2rem; + } + .manifesto { + font-size: 0.875rem; + } + .awareness-panel { + flex-direction: column; + align-items: flex-start; + } + .awareness-panel-img { + width: min(100%, 220px); + } + .api-grid { + grid-template-columns: 1fr; + } + .state-bar { + gap: 1.25rem; + flex-wrap: wrap; + } + .position-table { + min-width: 42rem; + } + .agents-table { + min-width: 38rem; + } + .position-table td:first-child { + width: 11rem; + } +} + +@media (max-width: 480px) { + .hero { + padding-top: 3rem; + } + .hero-tagline { + font-size: 1.7rem; + } + .manifesto { + font-size: 0.875rem; + } +} diff --git a/website/index.html b/website/index.html index f73c66ec..e639d355 100644 --- a/website/index.html +++ b/website/index.html @@ -1,978 +1,702 @@ - + - - - -
- supervise multiple coding agents in one terminal.
-run Claude Code, Codex, pi, opencode, droid, amp, and plain shells side by side in your existing terminal. see which agent is blocked, done, or still working without leaving the command line.
-workspaces, tabs, panes, automatic agent detection, and notifications. no electron, no web dashboard, no separate gui window.
-curl -fsSL https://herdr.dev/install.sh | sh
-
- supervise, triage, act
-orchestrate, spawn, wait
-the sidebar is split into two layers. scan the workspace list for the most urgent signal, then drill into the agents causing it in the selected workspace or across all workspaces.
-
+
+
+
+
+ herdr
+ tmux for agents.
++ herdr is an agent multiplexer that lives in your + terminal. +
++ workspaces, tabs, panes. mouse-native: click, drag, split. + every agent at a glance: + blocked, + working, + done. detach and reattach, + agents keep running. no gui app, no electron, no mac-only + native wrapper. you see the agent's own terminal, not + someone's interpretation of it. +
+curl -fsSL https://herdr.dev/install.sh | sh
each workspace shows one aggregate dot — the most urgent state across all its agents. yellow means something is working. you scan the full list in a glance.
+ +| + | tmux | +gui managers | +herdr | +
|---|---|---|---|
| persistent sessions | +✓ | +— | ++ ✓ + | +
| detach / reattach | +✓ | +— | ++ ✓ + | +
| panes, tabs, workspaces | +✓ | +✓ | ++ ✓ + | +
| agent awareness | +— | +✓ | ++ ✓ + | +
| lives in your terminal | +✓ | +— | ++ ✓ + | +
| real terminal views | +✓ | +— | ++ ✓ + | +
| mouse-native | +— | +✓ | ++ ✓ + | +
| lightweight binary | +✓ | +— | ++ ✓ + | +
| agents can orchestrate | +? | +? | ++ ✓ + | +
+ start herdr on your desktop or server. run your agents, + split panes, do your work. press + ctrl+b q to detach. close your terminal, + close your laptop; your agents keep running. open a new + terminal, run herdr, you're back. same + session, same panes, same agents. +
++ need to check on your agents from your phone? just ssh in + and run herdr. any ssh client works. no app to download, no + account to create. +
+
+
+
+
+
+ + each workspace rolls up to its most urgent state. + scan the full list in a glance. +
+the bottom section shows which specific agents are running in the selected workspace, or across all workspaces when you switch the panel scope. pi is working, droid is idle.
+
+ + see which agents are running, blocked, or finished. + no checking panes one by one. +
+one binary, no dependencies. run herdr in your terminal. onboarding takes 10 seconds.
press n to create a workspace. it opens as a terminal context, labels itself from git repo or folder name.
start your agents in panes like you normally would. herdr detects them automatically — no hooks, no config.
++ not a gui window. not a web dashboard. not electron. + herdr runs inside whatever terminal you already + use. + a single rust binary, no dependencies, nothing to configure. + it even works inside tmux. +
+herdr detects agents by reading foreground process and terminal output patterns. zero config required.
-| agent | -idle / done | -working | -blocked | -
|---|---|---|---|
| pi | -✓ | -✓ | -partial | -
| claude code | -✓ | -✓ | -✓ | -
| codex | -✓ | -✓ | -✓ | -
| droid | -✓ | -✓ | -✓ | -
| amp | -✓ | -✓ | -partial | -
| opencode | -✓ | -✓ | -✓ | -
| gemini cli | -detected, not fully tested | -||
| cursor agent | -detected, not fully tested | -||
| cline | -detected, not fully tested | -||
| agent | +idle / done | +working | +blocked | +
|---|---|---|---|
| pi | +✓ | +✓ | +partial | +
| + claude code + | +✓ | +✓ | +✓ | +
| + codex + | +✓ | +✓ | +✓ | +
| droid | +✓ | +✓ | +✓ | +
| amp | +✓ | +✓ | +✓ | +
| + opencode + | +✓ | +✓ | +✓ | +
| + gemini cli + | ++ detected, not fully tested + | +||
| + cursor agent + | ++ detected, not fully tested + | +||
| + cline + | ++ detected, not fully tested + | +||
agents can use herdr too
-herdr exposes a local unix socket that agents can talk to directly. create workspaces and tabs, split panes, spawn agents, read output, and wait for state changes.
-the same surface is available as CLI commands (herdr workspace, herdr tab, herdr pane, herdr wait) and as a reusable agent skill.
# create a workspace with a label -herdr workspace create --cwd ~/project --label "api server" ++-+++ agents can use herdr too. the local + unix socket lets agents create workspaces, split panes, + spawn other agents, read output, and wait for state + changes. +
++ available as CLI commands or as a reusable agent skill. +
++ socket api → + agent skill → +++# create a workspace and tab +herdr workspace create --cwd ~/project --label "api" +herdr tab create --label "logs" -# split a pane and run a command +# split a pane and run herdr pane split 1-1 --direction right herdr pane run 1-2 "npm test" -# wait for an agent to finish +# wait for an agent herdr wait agent-status 1-1 --status done -# read pane output -herdr pane read 1-2 --source recent-unwrapped --lines 50+# read output +herdr pane read 1-2 --source recent --lines 50 +