release: v2.3.5 — LM Studio auto-detection fix + Windows terminal cleanup

Reported via Discord #help-chat on 2026-04-21 by djoks.exe:
"LU does not recognize the models".

Root cause: AppShell's post-onboarding local-backend detection only
pre-enabled an openai-compat provider when exactly one backend was
detected. With two or more (very common setup — Ollama + LM Studio
together) it opened the BackendSelector modal without pre-enabling
anything. If the user dismissed the modal, LM Studio stayed disabled
and its models never showed up in the chat dropdown — exactly the
"LU doesn't see my models" symptom.

Fix: always pre-enable the first non-Ollama detected backend, even
when multiple are running. The selector modal stays as an educational
picker so the user can change which one is primary. Live-verified
against a mock LM Studio endpoint with Ollama also running on the
rebuilt release binary (hash differs from v2.3.4).

Also rolled into this release:
- CREATE_NO_WINDOW on the remaining Windows subprocess spawns that
  flashed a console window (state.rs taskkill at LU shutdown, search.rs
  docker pull/run at SearXNG install). Audit shows 100% coverage on
  the Windows code path now.

Suite 2161 -> 2166 green (+5 regression tests for the backend-autoenable
fix). tsc clean. cargo clean. Drop-in upgrade from v2.3.4, no breaking
changes, no localStorage migration.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: PurpleDoubleD

CHANGELOG.md

@@ -2,6 +2,18 @@
 
 All notable changes to Locally Uncensored are documented here.
 
+## [2.3.5] - 2026-04-21
+
+### Fixed
+- **LM Studio (and other openai-compat backends) now show up when Ollama is also running** — `AppShell`'s post-onboarding detection only auto-enabled a backend when exactly one was detected. With two or more (the very common Ollama + LM Studio setup) it showed the `BackendSelector` modal but pre-enabled nothing. Users who dismissed the modal saw zero LM Studio models in the chat dropdown even though LM Studio was clearly running — looked from the outside like "LU doesn't recognize my models". Reported via Discord `#help-chat` on 2026-04-21. Fix: the first non-Ollama detected backend is always pre-enabled (Ollama is left untouched since it has its own provider slot); the selector stays as an educational picker so you can change which openai-compat backend is primary. Reproduced live with a mock LM Studio endpoint on port 1234 with Ollama also running, verified the fix against the same setup on the release binary. Five regression tests in `AppShell-backend-autoenable.test.ts`.
+- **No more terminal flashes on Windows when LU kills subprocesses** — two Windows-branch `Command::new` spawns were missing `CREATE_NO_WINDOW`: the `taskkill` calls in `AppState::Drop` that tear down ComfyUI + Claude Code process trees on LU shutdown, and the `docker pull` / `docker run` in `search.rs` that installs SearXNG. Both briefly flashed a console window at the user. Now 100% of Windows-branch subprocess spawns carry the flag. LU itself never spawns LM Studio (only talks HTTP to a user-run instance), so the "no terminal when using LM Studio" guarantee was already true on that path; this tightens the peripheral surface.
+
+### Changed
+- Test suite 2161 → 2166 green (+5 regression tests for the backend-autoenable fix).
+
+### Notes
+- Drop-in upgrade from v2.3.4. No breaking changes. No localStorage migration. Everything from v2.3.4 (chat-history persistence, Ollama 0.21 compat, Codex loop guard, stop-button fast-path, stale-chip fix, 12-backend auto-detect, Mobile Remote, Codex streaming, Agent Mode rewrite, ERNIE-Image, Qwen 3.6, 75+ one-click model downloads) still applies.
+
 ## [2.3.4] - 2026-04-20
 
 ### Fixed

README.md

@@ -35,19 +35,16 @@ No cloud. No data collection. No API keys. Auto-detects 12 local backends. Your
 
 ---
 
-## v2.3.4 — Current Release
+## v2.3.5 — Current Release
 
-**Chat-history persistence fix, Ollama 0.21 compatibility, Codex loop guard, 2161 Tests**
+**LM Studio auto-detection fix + Windows terminal-popup cleanup, 2166 Tests**
 
 ### Critical Fixes (why you want this update)
-- **Chat history now survives updates** — NSIS auto-update, crashes, and abrupt process kills no longer wipe your conversations. `isTauri()` detection was broken in the v1→v2 migration, causing every backup/restore call to silently no-op. Fixed with dual-global detection + async-init polling + 5 s backup interval + event-driven + graceful-quit flush. Fully live-verified: destructive wipe + restore roundtrip confirmed on the release binary.
-- **Ollama 0.21 / 0.20.7 compatibility** — the auto-upgraded Ollama now rejects pre-existing models with `HTTP 404 model not found` on `/api/show` when their on-disk manifest lacks the new `capabilities` field. New top-of-app banner + Header Lichtschalter chip detect stale models and offer a one-click re-pull that verifies the fix before clearing the warning.
-- **Codex infinite-loop guard** — small 3 B coder models could get stuck repeating the same `file_write + shell_execute` batch forever when a test failed. Codex now halts with a clear "same tool sequence repeated — try a larger model" message.
-- **Stop button now instant** — abort signal checked between chunks in the for-await and NDJSON-reader loops. Thinking tokens no longer leak for 30–60 s after you click Stop on a Gemma-4 thinking response.
-- **Stale-chip state-leak** — switching from a stale model to a fresh one now clears the red toggle and chip immediately.
+- **LU now recognizes your LM Studio models when Ollama is also running** — if the first-launch detection found 2+ local backends (the very common "Ollama + LM Studio" setup), the backend selector modal opened but no provider got auto-enabled. Users who dismissed the modal saw zero LM Studio models in the dropdown even though LM Studio was clearly running. Fixed: the first non-Ollama backend is always pre-enabled; the selector stays as an educational picker so you can still switch primaries. Reproduced live and verified against a real LM Studio-like endpoint.
+- **No more terminal flashes on Windows** — a couple of subprocess spawns on the Windows code path were missing `CREATE_NO_WINDOW`, so killing ComfyUI/Claude Code during LU shutdown or installing SearXNG briefly flashed a console window. 100% of Windows-branch spawns are now flagged.
 
-### What's still in v2.3.4 from v2.3.3
-This is a hotfix release — v2.3.3's feature surface is unchanged.
+### What's still in v2.3.5 from v2.3.4
+This is a hotfix release — v2.3.4's feature surface (chat-history persistence, Ollama 0.21 compat, Codex loop guard, stop-button instant, stale-chip fix, 12 backend auto-detection, Mobile Remote, Codex streaming, Agent Mode, ERNIE-Image, Qwen 3.6, 75+ downloadable models) is unchanged. Every fix from v2.3.4 and earlier still applies.
 
 ### Remote Access + Mobile Web App
 - **Access your AI from your phone** — Dispatch via LAN or Cloudflare Tunnel (Internet)
@@ -78,7 +75,7 @@ This is a hotfix release — v2.3.3's feature surface is unchanged.
 - **AE-style text header** — clean typography for better discoverability
 - **Plugins dropdown** — Caveman Mode + Personas in one menu
 - **Thinking mode** — tri-state, auto-retry, universal tag stripper
-- **2161 tests** — comprehensive smoke tests covering the entire app
+- **2166 tests** — comprehensive smoke tests covering the entire app
 
 ---
 

docs/index.html

@@ -33,7 +33,7 @@
 <link rel="icon" href="https://locallyuncensored.com/favicon.png">
 <link rel="apple-touch-icon" href="https://locallyuncensored.com/favicon.png">
 <script type="application/ld+json">
-{"@context":"https://schema.org","@type":"SoftwareApplication","name":"Locally Uncensored","alternateName":["LU","Locally Uncensored Desktop App"],"applicationCategory":"DeveloperApplication","applicationSubCategory":"Artificial Intelligence","operatingSystem":"Windows 10, Windows 11, Linux","description":"Open-source desktop app for running AI locally. Chat with 20+ providers, generate images and video via ComfyUI, and code with a built-in AI agent — all on your own hardware. Auto-detects 12 local backends. No cloud, no telemetry.","url":"https://locallyuncensored.com/","downloadUrl":"https://github.com/PurpleDoubleD/locally-uncensored/releases/latest","softwareVersion":"2.3.4","license":"https://www.gnu.org/licenses/agpl-3.0.html","author":{"@type":"Person","name":"PurpleDoubleD","url":"https://github.com/PurpleDoubleD"},"offers":{"@type":"Offer","price":"0","priceCurrency":"USD","availability":"https://schema.org/InStock"},"featureList":["Plug & Play Setup with 12 Local Backend Auto-Detection (Ollama, LM Studio, vLLM, KoboldCpp, Jan, llama.cpp, LocalAI, GPT4All, TabbyAPI, Aphrodite, SGLang, TGI)","20+ Provider Presets (local + cloud)","Qwen 3.6 Day-0 Support (35B MoE, Vision, Agentic Coding, 256K context)","GPT-OSS Compatible (120B, 20B via Ollama)","GLM-4.7 Flash Integration","DeepSeek R1 and DeepSeek V3 Support","Llama 4 and Llama 3.3 Ready","Gemma 4 E4B and 27B (Native Vision)","Mistral Small 3 and Phi 4 Support","Qwen 3 Coder and Qwen 2.5 Coder","Abliterated Model Variants","Codex Coding Agent with Live Streaming and Apply-Patch","Claude Code CLI Integration","Agent Mode with 14 Tools, Parallel Execution, MCP, Sub-Agent Delegation","Remote Access from Phone via LAN or Cloudflare Tunnel (6-digit Passcode, QR Setup)","Mobile Web App (Chat, Codex, Tools, Plugins)","Image Generation via ComfyUI (FLUX 2 Klein, FLUX.1, Juggernaut XL, Z-Image Turbo, ERNIE-Image, SDXL, SD 3.5)","Video Generation (Wan 2.1, Wan 2.2, HunyuanVideo 1.5, LTX 2.3, FramePack F1, AnimateDiff, CogVideoX)","Image-to-Image with Denoise Control","Image-to-Video (FramePack F1, CogVideoX, SVD)","Granular Permission System (7 Tool Categories)","File Upload with Vision Support","Thinking Mode with Universal Tag Stripping","A/B Model Compare","Local Benchmark","Memory System and Document RAG","Voice (STT / TTS)","Auto-Update over Signed NSIS Channel","75+ One-Click Model Downloads","Hardware-Aware Model Recommendations","100% Offline and Private (No Telemetry)","AGPL-3.0 Open Source"],"screenshot":"https://raw.githubusercontent.com/PurpleDoubleD/locally-uncensored/master/docs/social-preview.png","softwareRequirements":"Any OpenAI-compatible local backend (Ollama, LM Studio, vLLM, KoboldCpp, etc.) or cloud API key","memoryRequirements":"8 GB RAM minimum; 8-16 GB VRAM for image/video generation","storageRequirements":"6 GB for default model"}
+{"@context":"https://schema.org","@type":"SoftwareApplication","name":"Locally Uncensored","alternateName":["LU","Locally Uncensored Desktop App"],"applicationCategory":"DeveloperApplication","applicationSubCategory":"Artificial Intelligence","operatingSystem":"Windows 10, Windows 11, Linux","description":"Open-source desktop app for running AI locally. Chat with 20+ providers, generate images and video via ComfyUI, and code with a built-in AI agent — all on your own hardware. Auto-detects 12 local backends. No cloud, no telemetry.","url":"https://locallyuncensored.com/","downloadUrl":"https://github.com/PurpleDoubleD/locally-uncensored/releases/latest","softwareVersion":"2.3.5","license":"https://www.gnu.org/licenses/agpl-3.0.html","author":{"@type":"Person","name":"PurpleDoubleD","url":"https://github.com/PurpleDoubleD"},"offers":{"@type":"Offer","price":"0","priceCurrency":"USD","availability":"https://schema.org/InStock"},"featureList":["Plug & Play Setup with 12 Local Backend Auto-Detection (Ollama, LM Studio, vLLM, KoboldCpp, Jan, llama.cpp, LocalAI, GPT4All, TabbyAPI, Aphrodite, SGLang, TGI)","20+ Provider Presets (local + cloud)","Qwen 3.6 Day-0 Support (35B MoE, Vision, Agentic Coding, 256K context)","GPT-OSS Compatible (120B, 20B via Ollama)","GLM-4.7 Flash Integration","DeepSeek R1 and DeepSeek V3 Support","Llama 4 and Llama 3.3 Ready","Gemma 4 E4B and 27B (Native Vision)","Mistral Small 3 and Phi 4 Support","Qwen 3 Coder and Qwen 2.5 Coder","Abliterated Model Variants","Codex Coding Agent with Live Streaming and Apply-Patch","Claude Code CLI Integration","Agent Mode with 14 Tools, Parallel Execution, MCP, Sub-Agent Delegation","Remote Access from Phone via LAN or Cloudflare Tunnel (6-digit Passcode, QR Setup)","Mobile Web App (Chat, Codex, Tools, Plugins)","Image Generation via ComfyUI (FLUX 2 Klein, FLUX.1, Juggernaut XL, Z-Image Turbo, ERNIE-Image, SDXL, SD 3.5)","Video Generation (Wan 2.1, Wan 2.2, HunyuanVideo 1.5, LTX 2.3, FramePack F1, AnimateDiff, CogVideoX)","Image-to-Image with Denoise Control","Image-to-Video (FramePack F1, CogVideoX, SVD)","Granular Permission System (7 Tool Categories)","File Upload with Vision Support","Thinking Mode with Universal Tag Stripping","A/B Model Compare","Local Benchmark","Memory System and Document RAG","Voice (STT / TTS)","Auto-Update over Signed NSIS Channel","75+ One-Click Model Downloads","Hardware-Aware Model Recommendations","100% Offline and Private (No Telemetry)","AGPL-3.0 Open Source"],"screenshot":"https://raw.githubusercontent.com/PurpleDoubleD/locally-uncensored/master/docs/social-preview.png","softwareRequirements":"Any OpenAI-compatible local backend (Ollama, LM Studio, vLLM, KoboldCpp, etc.) or cloud API key","memoryRequirements":"8 GB RAM minimum; 8-16 GB VRAM for image/video generation","storageRequirements":"6 GB for default model"}
 </script>
 <script type="application/ld+json">
 {"@context":"https://schema.org","@type":"Organization","name":"PurpleDoubleD","url":"https://github.com/PurpleDoubleD","logo":"https://locallyuncensored.com/logos/LU-monogram-bw-transparent.png","sameAs":["https://github.com/PurpleDoubleD","https://github.com/PurpleDoubleD/locally-uncensored","https://reddit.com/user/GroundbreakingMall54","https://locallyuncensored.com"]}
@@ -827,7 +827,7 @@
 
   <!-- Top meta strip -->
   <div class="mast-strip">
-    <span class="numero">v 2.3.4</span>
+    <span class="numero">v 2.3.5</span>
     <span class="sep">·</span>
     <span>AGPL-3.0</span>
     <span class="sep">·</span>
@@ -868,10 +868,10 @@ <h1 class="mast-title">Locally&nbsp;<span class="accent">Uncensored</span></h1>
   <div class="container">
     <div class="dl-card">
       <div class="dl-info">
-        <h3>Locally Uncensored <em>&mdash; 2.3.4</em></h3>
+        <h3>Locally Uncensored <em>&mdash; 2.3.5</em></h3>
         <p>WINDOWS &middot; ~152 MB &middot; AGPL-3.0</p>
       </div>
-      <a class="dl-btn" href="https://github.com/PurpleDoubleD/locally-uncensored/releases/download/v2.3.4/Locally.Uncensored_2.3.4_x64-setup.exe" download>
+      <a class="dl-btn" href="https://github.com/PurpleDoubleD/locally-uncensored/releases/download/v2.3.5/Locally.Uncensored_2.3.5_x64-setup.exe" download>
         <svg width="14" height="14" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.4" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M21 15v4a2 2 0 0 1-2 2H5a2 2 0 0 1-2-2v-4"/><polyline points="7 10 12 15 17 10"/><line x1="12" y1="15" x2="12" y2="3"/></svg>
         Download .exe
       </a>
@@ -1153,10 +1153,10 @@ <h3>Can I use this on macOS?</h3>
   <div class="narrow">
     <div class="dl-card">
       <div class="dl-info">
-        <h3>Locally Uncensored <em>&mdash; 2.3.4</em></h3>
+        <h3>Locally Uncensored <em>&mdash; 2.3.5</em></h3>
         <p>WINDOWS &middot; ~152 MB &middot; AGPL-3.0</p>
       </div>
-      <a class="dl-btn" href="https://github.com/PurpleDoubleD/locally-uncensored/releases/download/v2.3.4/Locally.Uncensored_2.3.4_x64-setup.exe" download>
+      <a class="dl-btn" href="https://github.com/PurpleDoubleD/locally-uncensored/releases/download/v2.3.5/Locally.Uncensored_2.3.5_x64-setup.exe" download>
         <svg width="14" height="14" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.4" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M21 15v4a2 2 0 0 1-2 2H5a2 2 0 0 1-2-2v-4"/><polyline points="7 10 12 15 17 10"/><line x1="12" y1="15" x2="12" y2="3"/></svg>
         Download .exe
       </a>

package.json

@@ -1,6 +1,6 @@
 {
   "name": "locally-uncensored",
-  "version": "2.3.4",
+  "version": "2.3.5",
   "private": false,
   "description": "Generate anything — text, images, video. Locally. Uncensored.",
   "license": "AGPL-3.0",

src-tauri/Cargo.lock

@@ -1,6 +1,6 @@
 [package]
 name = "locally-uncensored"
-version = "2.3.4"
+version = "2.3.5"
 description = "Private, local AI chat & image/video generation"
 authors = ["purpledoubled"]
 edition = "2021"

src-tauri/Cargo.toml

@@ -1,7 +1,7 @@
 {
   "$schema": "https://raw.githubusercontent.com/tauri-apps/tauri/dev/crates/tauri-cli/schema.json",
   "productName": "Locally Uncensored",
-  "version": "2.3.4",
+  "version": "2.3.5",
   "identifier": "com.purpledoubled.locally-uncensored",
   "build": {
     "beforeBuildCommand": "npm run build",