release: v1.7.4 — Engines/Library/Developer/Settings redesign + mobile responsiveness

README.md

@@ -177,8 +177,8 @@ the detail:
 - **Saved per-model profiles, per engine** — tune once per (model, engine) pair, so
   switching engines (or between two installs of the same engine, e.g. a fork) never
   overwrites another engine's tuning for the same model.
-- **Configurable VRAM headroom** (Settings → Engine, 300 MB–2 GB, default 1 GB) — tell auto-tune
-  how much VRAM to keep free for other GPU workloads instead of a fixed margin.
+- **Configurable VRAM headroom** (Settings → Models & loading → Advanced, 300 MB–2 GB, default
+  1 GB) — tell auto-tune how much VRAM to keep free for other GPU workloads instead of a fixed margin.
 
 </details>
 
@@ -216,7 +216,7 @@ the detail:
   in every chat with no restart. A **Research** persona forces multi-step web search and cites sources inline.
 - **Tool-call approval gate** — every tool call asks for your approval by default before it runs,
   with **Deny**, **Allow**, **Allow for this chat**, or **Always Allow** on an inline bar above the
-  composer. Set per-tool defaults globally from Developer → Tool permissions.
+  composer. Set per-tool defaults globally from Settings → Tools & safety.
 
 </details>
 
@@ -255,7 +255,7 @@ curl http://127.0.0.1:6996/v1/chat/completions \
 - **Anthropic-compatible** `/v1/messages` — including **tool use and streaming** — which powers
   Claude Code below. No other local host offers this.
 - **Structured output** — constrain any response to a **GBNF grammar** (or JSON shape).
-- **API-key auth** you can require when sharing over a LAN (Settings → Network).
+- **API-key auth** you can require when sharing over a LAN (Settings → Network & sharing).
 
 **The gateway loads models for you.** Most local hosts make you load a model first, then call it.
 TurboLLM's gateway reads the `model` field of any incoming request, **fuzzy-matches it to your
@@ -267,7 +267,7 @@ coding model, a vision model, and an embedder just names each one and it works 
 `opencode`, `kilo`, `openclaw`, and `hermes` — each gets pointed at TurboLLM the way that tool
 expects (config file merge, or its own CLI command) instead of a manual copy-paste setup. Inside
 Claude Code, `/model` lists your local models directly (with **Auto Model Swap** on in Settings →
-Gateway) so you can switch mid-session instead of only at launch.
+Models & loading) so you can switch mid-session instead of only at launch.
 
 </details>
 
@@ -285,7 +285,7 @@ fight for memory (and one usually OOMs). TurboLLM can hand the GPU over automati
 It's **push-based, not polling** — ComfyUI signals TurboLLM the moment a job starts/ends, so the
 handoff is immediate and deterministic (the model is gone *before* ComfyUI executes).
 
-**One-time setup** (Settings → ComfyUI): turn on **Pause for ComfyUI**, enter your ComfyUI folder
+**One-time setup** (Settings → Network & sharing → ComfyUI): turn on **Pause for ComfyUI**, enter your ComfyUI folder
 (the one containing `custom_nodes`), click **Install gate** (it writes a small custom node wired to
 this daemon), then **restart ComfyUI** once. The panel shows a live indicator (rendering / idle /
 connected); **Remove** undoes it.
@@ -333,7 +333,7 @@ daemon any time with **Ctrl+C**.
 No other local-LLM app lets you run **whatever inference engine you want**. TurboLLM treats
 the engine as a swappable component.
 
-**Add a custom engine** (Engines screen → **Add engine**):
+**Add a custom engine** (Engines screen → **Add your own engine**):
 
 1. Compile or download any `llama-server`-compatible binary — stock
    [llama.cpp](https://github.com/ggml-org/llama.cpp), a community fork, or your own build.
@@ -398,13 +398,14 @@ any OpenAI- or Anthropic-compatible tool (Open WebUI, Kilo Code, opencode, …).
 ## Use it from any device on your network
 
 The UI runs in the browser, so any phone, tablet, or laptop on your LAN can use the model on
-your GPU box:
+your GPU box — every screen is responsive down to phone widths, with the desktop layout
+untouched:
 
 ```bash
 turbollm --addr 0.0.0.0:6996    # bind all interfaces, then open http://<your-ip>:6996
 ```
 
-Turn on **Require API key** in Settings → Network when you expose it.
+Turn on **Require API key** in Settings → Network & sharing when you expose it.
 
 ---
 

turbollm/CHANGELOG.md

@@ -25,6 +25,51 @@ published version on npm has a matching `vX.Y.Z` tag in git.
 
 _Nothing yet._
 
+## [1.7.4] - 2026-07-07
+
+**Engines, Library, Developer, and Settings redesigned, and every screen now works on mobile.**
+
+### Added
+- Mobile-responsive layouts across every screen (drawer navigation, stacked cards, horizontal
+  tab strips, split panes that stack vertically) — the desktop/landscape layout is unchanged.
+- **Library**: a dedicated Folders dialog with a real folder picker for managing model
+  directories, plus name search and facet-aware filter chips (a chip only appears if at least
+  one model has it).
+- **Engines**: per-card hardware-fit indicators (green/amber/red: Compatible / runs after a
+  build / incompatible + reason) and grounded pros/cons for every engine.
+- The folder/file picker (used by model folders and "Add your own engine") gained an editable
+  address bar — paste or type an absolute path and jump there directly, alongside click-to-browse.
+
+### Changed
+- **Library**: rows are now a column-aligned two-tier layout (name + one state signal + quiet
+  capability chips, with aligned Quant/Size/Ctx/Speed columns); multi-quant models fold into one
+  row with a quant dropdown instead of a collapsible group.
+- **Engines**: the active-engine bar and hardware-fit selector merged into one panel with a
+  run-state traffic light (green running / amber starting / red error / grey stopped); llama.cpp
+  variants unified under one "Manage GPU builds" card; "Add your own engine" moved into a compact
+  strip.
+- **Developer**: narrowed to connecting external tools — a Connection panel, one-command CLI
+  setup cards, and a collapsed API reference. Cloud Deploy (RunPod) moved behind an internal
+  feature flag, off by default.
+- **Settings**: reorganized into a two-pane, five-category layout (General · Models & loading ·
+  Tools & safety · Network & sharing · System) with one sticky Save bar; expert knobs (GPU
+  layers, VRAM headroom, image/response token caps) moved into a collapsible Advanced section.
+  Tool permissions moved here from Developer (Settings → Tools & safety).
+- "Add your own engine" filesystem browsing is local-only (same restriction as engine scanning)
+  but can now browse the whole local filesystem, including other drives on Windows — previously
+  confined to your home directory.
+
+### Fixed
+- Every dialog now keeps a gutter on narrow screens instead of rendering flush to the viewport edge.
+- Assorted mobile overflow issues: chat header controls clipping on narrow screens, Library
+  action buttons pushed off-screen, Settings controls clipped at the edge, and long Hugging Face
+  repo names/inline code overflowing their containers.
+
+### Discord
+- New, clearer look for the Engines, Library, Developer, and Settings screens.
+- The whole app now works properly in your phone's browser, not just on desktop.
+- Picking a model folder or a custom engine now lets you paste in a path directly, or browse to any drive.
+
 ## [1.7.3] - 2026-07-06
 
 **The standalone Agents screen is retired — replaced by a shared Skills library used directly in chat, plus a new Agents tab in Customize for editing personas and building your own.**

turbollm/README.md

@@ -177,8 +177,8 @@ the detail:
 - **Saved per-model profiles, per engine** — tune once per (model, engine) pair, so
   switching engines (or between two installs of the same engine, e.g. a fork) never
   overwrites another engine's tuning for the same model.
-- **Configurable VRAM headroom** (Settings → Engine, 300 MB–2 GB, default 1 GB) — tell auto-tune
-  how much VRAM to keep free for other GPU workloads instead of a fixed margin.
+- **Configurable VRAM headroom** (Settings → Models & loading → Advanced, 300 MB–2 GB, default
+  1 GB) — tell auto-tune how much VRAM to keep free for other GPU workloads instead of a fixed margin.
 
 </details>
 
@@ -216,7 +216,7 @@ the detail:
   in every chat with no restart. A **Research** persona forces multi-step web search and cites sources inline.
 - **Tool-call approval gate** — every tool call asks for your approval by default before it runs,
   with **Deny**, **Allow**, **Allow for this chat**, or **Always Allow** on an inline bar above the
-  composer. Set per-tool defaults globally from Developer → Tool permissions.
+  composer. Set per-tool defaults globally from Settings → Tools & safety.
 
 </details>
 
@@ -255,7 +255,7 @@ curl http://127.0.0.1:6996/v1/chat/completions \
 - **Anthropic-compatible** `/v1/messages` — including **tool use and streaming** — which powers
   Claude Code below. No other local host offers this.
 - **Structured output** — constrain any response to a **GBNF grammar** (or JSON shape).
-- **API-key auth** you can require when sharing over a LAN (Settings → Network).
+- **API-key auth** you can require when sharing over a LAN (Settings → Network & sharing).
 
 **The gateway loads models for you.** Most local hosts make you load a model first, then call it.
 TurboLLM's gateway reads the `model` field of any incoming request, **fuzzy-matches it to your
@@ -267,7 +267,7 @@ coding model, a vision model, and an embedder just names each one and it works 
 `opencode`, `kilo`, `openclaw`, and `hermes` — each gets pointed at TurboLLM the way that tool
 expects (config file merge, or its own CLI command) instead of a manual copy-paste setup. Inside
 Claude Code, `/model` lists your local models directly (with **Auto Model Swap** on in Settings →
-Gateway) so you can switch mid-session instead of only at launch.
+Models & loading) so you can switch mid-session instead of only at launch.
 
 </details>
 
@@ -285,7 +285,7 @@ fight for memory (and one usually OOMs). TurboLLM can hand the GPU over automati
 It's **push-based, not polling** — ComfyUI signals TurboLLM the moment a job starts/ends, so the
 handoff is immediate and deterministic (the model is gone *before* ComfyUI executes).
 
-**One-time setup** (Settings → ComfyUI): turn on **Pause for ComfyUI**, enter your ComfyUI folder
+**One-time setup** (Settings → Network & sharing → ComfyUI): turn on **Pause for ComfyUI**, enter your ComfyUI folder
 (the one containing `custom_nodes`), click **Install gate** (it writes a small custom node wired to
 this daemon), then **restart ComfyUI** once. The panel shows a live indicator (rendering / idle /
 connected); **Remove** undoes it.
@@ -333,7 +333,7 @@ daemon any time with **Ctrl+C**.
 No other local-LLM app lets you run **whatever inference engine you want**. TurboLLM treats
 the engine as a swappable component.
 
-**Add a custom engine** (Engines screen → **Add engine**):
+**Add a custom engine** (Engines screen → **Add your own engine**):
 
 1. Compile or download any `llama-server`-compatible binary — stock
    [llama.cpp](https://github.com/ggml-org/llama.cpp), a community fork, or your own build.
@@ -398,13 +398,14 @@ any OpenAI- or Anthropic-compatible tool (Open WebUI, Kilo Code, opencode, …).
 ## Use it from any device on your network
 
 The UI runs in the browser, so any phone, tablet, or laptop on your LAN can use the model on
-your GPU box:
+your GPU box — every screen is responsive down to phone widths, with the desktop layout
+untouched:
 
 ```bash
 turbollm --addr 0.0.0.0:6996    # bind all interfaces, then open http://<your-ip>:6996
 ```
 
-Turn on **Require API key** in Settings → Network when you expose it.
+Turn on **Require API key** in Settings → Network & sharing when you expose it.
 
 ---
 

turbollm/package.json

@@ -1,6 +1,6 @@
 {
   "name": "turbollm",
-  "version": "1.7.3",
+  "version": "1.7.4",
   "description": "TurboLLM — local LLM platform: run any inference engine auto-tuned to your GPU, with a web UI and OpenAI/Anthropic-compatible API. Point Claude Code at your own machine in one command.",
   "license": "FSL-1.1-ALv2",
   "author": "Mohit Soni",

turbollm/src/api/routes.ts

@@ -2,7 +2,7 @@
 import type { Context, Hono } from 'hono'
 import { streamSSE } from 'hono/streaming'
 import { existsSync, mkdirSync, readFileSync, readdirSync, realpathSync, rmSync, statSync, writeFileSync } from 'node:fs'
-import { basename, dirname, join, resolve, sep } from 'node:path'
+import { basename, dirname, join, resolve } from 'node:path'
 import { GATE_VERSION, gateNodeSource } from '../comfyui/gate-template'
 import { randomUUID } from 'node:crypto'
 import { homedir, networkInterfaces } from 'node:os'
@@ -944,26 +944,31 @@ export function registerApi(app: Hono, d: Deps): void {
     }
   })
 
-  // ---- filesystem browser (spec 03 §9): pick an engine binary by navigating the
-  // disk from the browser. Loopback-only and confined to the user's home dir so a
-  // page on the LAN cannot read arbitrary files through the daemon.
+  // ---- filesystem browser (spec 03 §9): pick an engine binary/folder by navigating
+  // the disk from the browser. LOCAL-ONLY (isLocalRequest, matching /engines/scan): a
+  // page on the LAN — even with a valid key — cannot read the filesystem through the
+  // daemon; only the browser on the machine running TurboLLM can. Deliberately NOT
+  // home-confined, so an engine that lives on another drive (D:\) or outside home is
+  // reachable — the local user already picks (and executes) arbitrary binaries via scan.
   app.get('/api/v1/fs/browse', (c) => {
-    const home = realHome()
+    if (!isLocalRequest(c, d))
+      return err(c, 403, 'forbidden', 'Browsing the filesystem is only available on the machine running TurboLLM.')
     const raw = (c.req.query('path') ?? '').trim()
-    // Resolve the requested path; default to the home dir when none is given.
-    const target = raw ? resolve(raw) : home
-    // Canonicalize symlinks before the containment check so a symlink inside home
-    // that points outside cannot be used to escape. Fall back to the lexical path
-    // if the target doesn't exist yet (it then fails the readdir below cleanly).
+    // Windows "This PC": a virtual root above the drive letters, so the user can hop
+    // between drives (C:\, D:\, …). Reached by going "up" from a drive root.
+    if (raw === FS_DRIVES_ROOT && process.platform === 'win32') {
+      return c.json({ path: FS_DRIVES_ROOT, parent: null, entries: listWindowsDrives() })
+    }
+    // Default to the home dir (a sensible starting point); any path is allowed.
+    const target = raw ? resolve(raw) : homedir()
+    // Canonicalize symlinks; fall back to the lexical path if it doesn't exist yet
+    // (it then fails the readdir below cleanly).
     let real: string
     try {
       real = realpathSync(target)
     } catch {
       real = target
     }
-    if (!isWithinHome(real, home)) {
-      return err(c, 403, 'path_outside_home', 'That folder is outside your home directory.')
-    }
     let entries: { name: string; path: string; isDir: boolean }[]
     try {
       entries = readdirSync(real, { withFileTypes: true })
@@ -985,9 +990,11 @@ export function registerApi(app: Hono, d: Deps): void {
     } catch {
       return err(c, 400, 'fs_read_failed', 'Could not read that folder (permission denied or not a directory).')
     }
-    // Parent is null at the home root or once it would escape home.
-    const parentDir = real === home ? null : dirname(real)
-    const parent = parentDir && isWithinHome(parentDir, home) ? parentDir : null
+    // Parent: the containing dir; at a filesystem/drive root, go up to the Windows drive
+    // list (so the user can switch drives) or stop (null) on a single-root OS.
+    const parentDir = dirname(real)
+    const parent =
+      parentDir === real ? (process.platform === 'win32' ? FS_DRIVES_ROOT : null) : parentDir
     return c.json({ path: real, parent, entries })
   })
 
@@ -2236,24 +2243,23 @@ function readTail(path: string, n: number): string[] {
 
 // ── filesystem browser helpers (spec 03 §9) ─────────────────────────────────
 
-/** The user's home dir, canonicalized (symlinks resolved) so the containment
- *  check below compares like-with-like. */
-function realHome(): string {
-  const h = homedir()
-  try {
-    return realpathSync(h)
-  } catch {
-    return h
-  }
-}
+/** Sentinel path for the Windows "This PC" drive list — a virtual root above the
+ *  drive letters (C:\, D:\, …) so the browser can hop between drives. */
+const FS_DRIVES_ROOT = '::drives'
 
-/** True when `p` is the home dir itself or a descendant of it. Compares the
- *  normalized paths and requires a trailing separator on the prefix so
- *  `/home/bobby` is not treated as inside `/home/bob`. Cross-platform: `sep`
- *  is `\` on Windows, `/` elsewhere. */
-function isWithinHome(p: string, home: string): boolean {
-  if (p === home) return true
-  return p.startsWith(home.endsWith(sep) ? home : home + sep)
+/** Available Windows drive roots (C:\, D:\, …) as browsable entries. Probes A–Z;
+ *  only the drives that exist on the box respond, missing ones are skipped. */
+function listWindowsDrives(): { name: string; path: string; isDir: boolean }[] {
+  const drives: { name: string; path: string; isDir: boolean }[] = []
+  for (let i = 65; i <= 90; i++) {
+    const root = `${String.fromCharCode(i)}:\\`
+    try {
+      if (existsSync(root)) drives.push({ name: root, path: root, isDir: true })
+    } catch {
+      /* unreadable drive (e.g. an empty removable slot) — skip */
+    }
+  }
+  return drives
 }
 
 function getLanIp(): string {

turbollm/web/src/components/ModelLoadMenu.tsx

@@ -41,7 +41,7 @@ export function ModelLoadMenu({
   return (
     <DropdownMenu>
       <DropdownMenuTrigger
-        className="flex h-8 max-w-[260px] items-center gap-1.5 rounded-md border border-border bg-panel px-2.5 text-[13px] text-ink transition-colors hover:border-[color:var(--accent)] disabled:opacity-60"
+        className="flex h-8 max-w-[160px] items-center gap-1.5 rounded-md border border-border bg-panel px-2.5 text-[13px] text-ink transition-colors hover:border-[color:var(--accent)] disabled:opacity-60 md:max-w-[260px]"
         disabled={pending}
       >
         {pending || ejecting ? (

turbollm/web/src/components/common.tsx

@@ -67,8 +67,8 @@ export function ScreenHeader({
   actions?: ReactNode
 }) {
   return (
-    <div className="mb-6 flex items-start justify-between gap-4">
-      <div>
+    <div className="mb-6 flex flex-col gap-3 sm:flex-row sm:items-start sm:justify-between sm:gap-4">
+      <div className="min-w-0">
         <h1 className="text-[18px] font-semibold tracking-[-0.01em] text-ink">{title}</h1>
         {description && <p className="mt-1 text-[13px] text-muted">{description}</p>}
       </div>

turbollm/web/src/components/ui/alert-dialog.tsx

@@ -27,7 +27,7 @@ export const AlertDialogContent = React.forwardRef<
     <AlertDialogPrimitive.Content
       ref={ref}
       className={cn(
-        'fixed left-1/2 top-1/2 z-50 w-full max-w-md -translate-x-1/2 -translate-y-1/2',
+        'fixed left-1/2 top-1/2 z-50 w-[calc(100%_-_2rem)] max-w-md -translate-x-1/2 -translate-y-1/2',
         'rounded-[var(--radius-lg)] border border-border bg-panel p-5 shadow-[var(--shadow-2)]',
         'focus:outline-none',
         className,
@@ -49,7 +49,7 @@ export function AlertDialogFooter({
   className,
   ...props
 }: React.HTMLAttributes<HTMLDivElement>) {
-  return <div className={cn('mt-5 flex justify-end gap-2', className)} {...props} />
+  return <div className={cn('mt-5 flex flex-wrap justify-end gap-2', className)} {...props} />
 }
 
 export const AlertDialogTitle = React.forwardRef<