hermes-agent/apps/desktop
brooklyn! 3c163cb035
feat(desktop): background needs-input indicator, clarify redesign, Cmd+K palette & UI consistency pass (#38631)
* fix(desktop): surface background-session clarify prompts instead of hanging

clarify.request is a one-shot blocking event: the gateway turn blocks on
clarify.respond. The desktop handler dropped it for any non-focused session
(`if (!isActiveEvent) return`) and stored at most one request in a single
global atom, so a background session that asked a clarifying question hung
forever and re-focusing it could never recover (the event was already gone).

- store/clarify.ts: key pending requests by runtime session id; expose the
  active session's request via a focus-scoped computed view (ClarifyTool is
  unchanged). clearClarifyRequest takes an optional session id for targeted
  clears, with a request-id fallback.
- use-message-stream.ts: park every session's clarify (drop the isActiveEvent
  early return); toast when one lands for a background session since the row
  otherwise just keeps spinning like normal work.
- clarify-tool.tsx: clear by session id so answering one chat can't wipe
  another's pending request.
- store/clarify.test.ts: concurrent independence, focus-scoped view,
  targeted/stale/fallback clears.

* feat(desktop): persistent needs-input indicator + icon button consolidation

Replace the background-clarify toast (expired on alt-tab, easy to miss) with a
persistent, glowing amber "needs input" dot on the session's sidebar row,
driven off a new ClientSessionState.needsInput flag mirrored into a
$attentionSessionIds store. The flag is set on clarify.request and cleared the
moment the turn resumes (tool.complete) or ends.

Also: redesign the clarify tool UI (borderless choices, pseudo-radio dots,
right-aligned checkmark, arc border, tighter padding), make Button the single
source of icon-button styling (4px radius, new icon-titlebar variant, titlebar
buttons rendered polymorphically via asChild, Codicons throughout), put the
file-tree refresh action first, and .trim() pasted composer text.

* style(desktop): padding-driven, square non-icon buttons

Default button sizing was vanilla-shadcn chunky (fixed h-9, 16px padding) and
inconsistent with the icon-button radius pass. Size text variants by
padding + line-height instead of fixed heights so they stay snug and scale
with content, and drop the radius on non-icon buttons (icon buttons keep the
shared 4px). Move the update-overlay CTAs off a hardcoded h-10 onto the
padding-based lg variant. Composer and the inline approval strip are untouched.

* style(desktop): shrink button scale, flush overlay sidebar, variant-ize stray buttons

- Buttons: smaller default font (14px -> 13px) and tighter padding-driven sizes
  across every variant; the chunky shadcn scale read as oversized in a dense
  desktop UI.
- Overlay split layout (settings / command center): the shared OverlayView top
  padding left the card surface showing as a gap above the sidebar. Move the
  titlebar clearance into each column so the sidebar background runs flush to
  the card's top edge.
- Consolidate buttons that hardcoded size/radius/font onto the proper size
  variants (tooltip-icon-button, overlay close, cron IconAction, SidebarTrigger,
  gateway system button, session-row actions radius, title chip radius, release
  notes link) so styling flows from variant props, not per-call overrides.
  Composer and the inline approval strip are intentionally left as-is.

* style(desktop): 12px button text, drop sparkle decoration + redundant settings titles

- Button base font down to 12px (text-xs) for the dense desktop scale.
- Remove the decorative Sparkles glyph from the model "Apply" button (keep the
  spinner while applying).
- Drop the page-level section titles that just restate the left nav ("Main
  model", "Appearance", "MCP servers") — the sidebar already labels the pane.
  Sub-section headings (Auxiliary models, LLM providers, etc.) stay.

* feat(desktop): add boxless `text` button variant; use for aux-model actions

New reusable `text` variant renders a button as inline label text (no
bg/border, muted -> foreground, underline-on-hover affordance). Emphasize the
actionable word by adding `font-semibold`/`underline` at the call site. Applied
to the auxiliary-model "Set to main" (plain), "Change" and "Reset all to main"
(bold + underlined) actions, replacing the boxed ghost/outline buttons.

* style(desktop): nudge button scale up + 2.5px radius on non-icon buttons

Bump default/sm vertical padding a step (the 12px pass read too small) and give
non-icon buttons a subtle 2.5px radius instead of square corners. Icon buttons
keep their 4px.

* style(desktop): unify Input/Textarea/SelectTrigger on shared controlVariants

Mirror the buttonVariants exercise for non-composer form controls: add a
single controlVariants source of truth (2.5px radius, 12px text,
padding-driven sizing, chrome via desktop-input-chrome) and consume it from
Input, Textarea, and SelectTrigger. Drop per-call radius/height/font
overrides that fought the shared look.

* style(desktop): flatten appearance settings — drop card-in-card sections

Remove the outer card chrome (border/bg/shadow/rounded) wrapping each
appearance section so they're flat headings + option grids instead of
boxes nested inside boxes, matching the other settings pages.

* style(desktop): de-box appearance options into flat rows + bare theme swatches

Color Mode and Tool Call Display become flat radio-style rows (no tile
border/fill, no inner icon box, no filled check badge — just a subtle active
bg and a check). Theme drops its outer card wrapper so only the preview
swatch shows, with a primary ring marking the active palette.

* style(desktop): primitive-level pointer cursor + borderless settings lists

Add a base-layer rule giving every interactive control (button, select,
menu item, switch, tab, summary) cursor:pointer, and strip the now-redundant
hardcoded cursor-pointer from those elements (plain clickable divs/labels
keep theirs). Remove the divide-y separators from settings list sections so
they breathe.

* style(desktop): Color Mode + Tool Call Display as one-row segmented controls

Replace the vertical option-row lists with a compact SegmentedControl
(grouped pill buttons on a single track), dropping the per-option
descriptions since the section subtitle already covers the context.

* style(desktop): drop redundant On/Off label next to boolean config switches

The switch already communicates state, so the text label was noise.

* style(desktop): add Switch xs size; move appearance controls inline-right

Add an xs size variant to the Switch primitive and use it for the provider
edit submenu toggles. In appearance settings, drop the redundant selection
Pills (the UI already shows the active choice), move the Color Mode and Tool
Call Display segmented controls into the section header's right side
(responsive: stacks under the heading on narrow widths), and shrink the
segmented control.

* feat(desktop): titlebar toggle to flip sidebar sides

Adds a top-left swap button (replacing the search icon) that mirrors the
layout: sessions sidebar ↔ file browser + preview rail. Persisted via
$panesFlipped. The left/right sidebar toggles, content inset, and pane
borders all follow the active side so the buttons stay accurate after a flip.

* feat(desktop): global Cmd+K palette + UI consistency overhaul

Builds on the clarify/needs-input work with a cross-cutting pass to make
the desktop surfaces feel like one app.

- Global Cmd+K command palette (cmdk): nav, settings deep-links, async
  API-key / MCP-server / archived-session groups, reusable theme sub-page
  (light/dark groups, stays open on pick), loop nav, fuzzy match. Replaces
  per-page settings search.
- Shared SearchField: borderless, underline-on-focus, `field-sizing`
  auto-width. Unifies sessions sidebar, pages, overlays, command center,
  cron; drops bespoke OverlaySearchInput.
- Cron & Profiles converted to OverlayView; flat token-driven panels
  (no card-in-card / divider borders) matching command center.
- `r` refresh hotkey via useRefreshHotkey; drop the visible refresh buttons.
- Button text/textStrong link variants applied across settings & views;
  shared PAGE_INSET_X content gutters.
- Math/ascii loaders replace "Loading…" text placeholders; x-icon close
  over text "Close"; cursor-pointer at the dropdown/select primitive level.

* style(desktop): tidy root error-boundary actions

Reload window → text link, Open logs pushed right (ml-auto), and the
error message box drops the oversized rounded-2xl for rounded-md.

* style(desktop): fix profiles sidebar — header + add-icon, drop text-link

The full-width `text` New-profile button drew an underline under the +
glyph on hover (text-decoration spans the icon). Replace with a proper
"PROFILES" section header + ghost add-icon button, matching the chat
sidebar's header/new-item pattern.

* style(desktop): kill focus rings globally

Tab/focus showed Tailwind's `focus-visible:ring-*` (a box-shadow) plus the
native outline. Drop both via an unlayered reset that nulls --tw-ring-*;
the composer / input soft-glow is untouched (those use direct box-shadows).

* style(desktop): shared Badge component; tidy profile metadata

Add a proper shadcn-style Badge (CVA tones, app radius — not a full pill)
and use it for the Default/.env tags instead of bespoke rounded-full spans.
Drop the oversized text-sm metadata values to text-xs.

* style(desktop): migrate bespoke pills to shared Badge; tidy cron/titlebar

- Sidebar toggles in the titlebar no longer carry an active highlight —
  they're plain show/hide affordances now.
- Replace every bespoke rounded-full status pill (cron, messaging,
  settings, skills) with the shared Badge (adds a `warn` tone). App radius,
  one component.
- Cron row actions use Codicons (play/debug-pause/zap/edit/trash) to match
  the rest of the chrome instead of stray lucide glyphs.

* style(desktop): drop active background on titlebar actions

Mute/haptics state reads from the icon glyph (and aria-pressed) — no
background highlight on any titlebar action.

* style(desktop): tighten error-boundary action gap

gap-4 → gap-2.5 between Try again / Reload window.

* style(desktop): hide search when there's nothing to search

Empty datasets no longer render a search field. Adds a `searchHidden` prop
to PageSearchShell (artifacts/skills/messaging) and gates cron + command
center sessions search on a non-empty list. The chat sidebar already did
this via showSessionSections.

* fix(desktop): composer wraps long text & expands at the real wrap point

Long unbroken input ran off horizontally and the stacked layout flipped
on a char-count guess (too early). Add wrap rules to the contentEditable
and drive expansion off the editor's actual rendered height via the
resize observer, so it stacks exactly when the text wraps to a 2nd line.

* feat(desktop): composer/intro polish + shared ErrorState

- Composer single-line row centers (was bottom-aligned); placeholder
  randomizes per session (starter vs follow-up) without mid-stream flip.
- Drop chat header on brand-new sessions (dead label + border).
- ⌘N flashes its sidebar hint; ⌘. toggles the command center.
- Intro wordmark fills width (drop 8rem fit cap).
- Unify error states on a shared ErrorState component (boundary + updates).

* style(desktop): satisfy lint across PR-touched files

* refactor(desktop): DRY/elegance pass over PR-touched files

- Shared useDeepLinkHighlight hook collapses 3 near-identical settings
  deep-link effects (keys/mcp); config kept inline (distinct bail-clear).
- command-center: table-driven SECTION_ICONS + single errorText helper.
- clarify-tool: OPTION_ROW_CLASS + RadioDot extracted from option rows.
- desktop-controller: merge Cmd+K / Cmd+. into one keydown handler.
- statusbar-controls: hoist shared action class.
- Misc: drop redundant cn()/cursor-pointer/dead fields; tidy switch.

* feat(desktop): Cmd+K jumps to sessions; drop API-key entries

Add active sessions to the palette (fuzzy jump-to-chat), remove the
low-value per-API-key entries, and move the lazy palette sources
(config/sessions/archived) to react-query instead of hand-rolled
useState + effect fetching. Hoist the shared nav helper.
2026-06-04 00:47:08 -05:00
..
assets Add Hermes desktop app (#20059) 2026-05-31 17:46:56 -05:00
electron fix(desktop): attachments on Enter, IME composition, scroll, fetchJson resets (salvage #38502) (#38677) 2026-06-03 23:38:58 -05:00
public Add Hermes desktop app (#20059) 2026-05-31 17:46:56 -05:00
scripts fix(desktop): stop chat scroll jumping by disabling native scroll anchoring 2026-06-02 23:08:01 -05:00
src feat(desktop): background needs-input indicator, clarify redesign, Cmd+K palette & UI consistency pass (#38631) 2026-06-04 00:47:08 -05:00
.prettierrc Add Hermes desktop app (#20059) 2026-05-31 17:46:56 -05:00
components.json Add Hermes desktop app (#20059) 2026-05-31 17:46:56 -05:00
eslint.config.mjs Add Hermes desktop app (#20059) 2026-05-31 17:46:56 -05:00
index.html fix(desktop): triage batch of GUI quality-of-life fixes (#37536) 2026-06-02 16:33:22 -04:00
package.json fix(install): require Node >=20.19/22.12 for the desktop build 2026-06-03 09:19:04 -05:00
preview-demo.html Add Hermes desktop app (#20059) 2026-05-31 17:46:56 -05:00
README.md fix(desktop): triage batch of GUI quality-of-life fixes (#37536) 2026-06-02 16:33:22 -04:00
tsconfig.json Add Hermes desktop app (#20059) 2026-05-31 17:46:56 -05:00
vite.config.ts Add Hermes desktop app (#20059) 2026-05-31 17:46:56 -05:00

Hermes Desktop ☤

Download Documentation Discord License: MIT

The native desktop app for Hermes Agent — the self-improving AI agent from Nous Research. Same agent, same skills, same memory as the CLI and gateway, in a polished native window — chat with streaming tool output, side-by-side previews, a file browser, voice, and settings, no terminal required. Available for macOS, Windows, and Linux.

Chat with the full agentStreaming responses, live tool activity, structured tool summaries, and the same conversation history as every other Hermes surface.
Side-by-side previewsRender web pages, files, and tool outputs in a right-hand pane while you keep chatting.
File browserExplore and preview the working directory without leaving the app.
VoiceTalk to Hermes and hear it back.
Settings & onboardingManage providers, models, tools, and credentials from a real UI. First-run setup gets you to your first message in seconds.
Stays currentBuilt-in updates pull the latest agent and rebuild the app in place.

Install

Add --include-desktop to the one-line installer and it sets up the agent and builds the desktop app in one go:

curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash -s -- --include-desktop

Already have the Hermes CLI? Just run:

hermes desktop

It builds and launches the GUI against your existing install — same config, keys, sessions, and skills. On first launch Hermes walks you through picking a provider and model; nothing else to configure.

Prebuilt installers

When a release ships desktop installers they're attached to its releases page.dmg (macOS), .exe / .msi (Windows), .AppImage / .deb / .rpm (Linux). These are published manually, so the install-with-Hermes path above is the most reliable way to get the latest.


Updating

The app checks for updates in the background and offers a one-click update when one is ready. You can also update any time from the CLI:

hermes update

Requirements

The installer handles everything for you (Python 3.11+, a portable Git, ripgrep). The only thing worth knowing:

  • Windows — the installer bundles its own Git and Python; no admin rights or system changes required.
  • macOS / Linux — uses your system Python 3.11+ (installed automatically if missing).

Development

Want to hack on the app itself? Install workspace deps from the repo root once, then run the dev server from this directory:

npm install          # from repo root — links apps/desktop, web, apps/shared
cd apps/desktop
npm run dev          # Vite renderer + Electron, which boots the Python backend

Point the app at a specific source checkout, or sandbox it away from your real config:

HERMES_DESKTOP_HERMES_ROOT=/path/to/clone npm run dev
HERMES_HOME=/tmp/throwaway npm run dev
npm run dev:fake-boot   # exercise the startup overlay with deterministic delays

Building installers

npm run dist:mac     # DMG + zip
npm run dist:win     # NSIS + MSI
npm run dist:linux   # AppImage + deb + rpm
npm run pack         # unpacked app under release/ (no installer)

Installers are built and uploaded to GitHub Releases manually. macOS/Windows signing & notarization happen automatically when the relevant credentials are present in the environment (CSC_LINK / CSC_KEY_PASSWORD / APPLE_* for macOS, WIN_CSC_* for Windows).

How it works

The packaged app ships only the Electron shell. On first launch it installs the Hermes Agent runtime into HERMES_HOME (~/.hermes, or %LOCALAPPDATA%\hermes on Windows) — the same layout a CLI install uses, so the two are interchangeable. The renderer (React, in src/) talks to a hermes dashboard --tui backend over the standard gateway APIs and reuses the embedded TUI rather than reimplementing chat. The install, backend-resolution, and self-update logic all live in electron/main.cjs.

Verification

Run before opening a PR (lint may surface pre-existing warnings but must exit cleanly):

npm run fix
npm run type-check
npm run lint
npm run test:desktop:all

Troubleshooting

Boot logs land in HERMES_HOME/logs/desktop.log (includes backend output and recent Python tracebacks) — check it first if the app reports a boot failure.

macOS / Linux:

# Force a clean first-launch setup
rm "$HOME/.hermes/hermes-agent/.hermes-bootstrap-complete"
# Rebuild a broken Python venv
rm -rf "$HOME/.hermes/hermes-agent/venv"
# Reset a stuck macOS microphone prompt (macOS only)
tccutil reset Microphone com.nousresearch.hermes

Windows (PowerShell):

# Force a clean first-launch setup
Remove-Item "$env:LOCALAPPDATA\hermes\hermes-agent\.hermes-bootstrap-complete"
# Rebuild a broken Python venv
Remove-Item -Recurse -Force "$env:LOCALAPPDATA\hermes\hermes-agent\venv"

The default Hermes home on Windows is %LOCALAPPDATA%\hermes. Set the HERMES_HOME env var if you've relocated it.


Community


License

MIT — see LICENSE.

Built by Nous Research.