sync upstream/main with origin/main

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

Author: ofirgluzman

.claude/skills/video-editor-development/SKILL.md

@@ -0,0 +1,6 @@
+---
+name: video-editor-development
+description: Architecture guidance for the LTX video editor — editor store, selectors, actions, undo/redo, persistence, keyboard behavior, timeline semantics, and hot-path performance.
+---
+
+@../../docs/skills/video-editor-development.md

.codex/skills/ltx-video-editor-development/SKILL.md

@@ -0,0 +1,3 @@
+# LTX Video Editor Development
+
+Refer to [docs/skills/video-editor-development.md](../../../docs/skills/video-editor-development.md) for the full guidance.

.codex/skills/ltx-video-editor-development/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "LTX Video Editor Development"
+  short_description: "Guidance for editing the LTX Desktop video editor safely"
+  default_prompt: "Use $ltx-video-editor-development to implement or review changes in the LTX Desktop video editor."

.cursor/skills/video-editor-development/SKILL.md

@@ -0,0 +1,6 @@
+---
+name: video-editor-development
+description: Architecture guidance for the LTX video editor — editor store, selectors, actions, undo/redo, persistence, keyboard behavior, timeline semantics, and hot-path performance.
+---
+
+@../../../docs/skills/video-editor-development.md

.github/workflows/ci.yml

@@ -27,9 +27,6 @@ jobs:
 
       - run: pnpm install --frozen-lockfile
 
-      - name: TypeScript type check
-        run: pnpm run typecheck:ts
-
       - uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5.4.2
         with:
           version: "latest"
@@ -48,6 +45,12 @@ jobs:
         working-directory: backend
         run: uv run pyright
 
+      - name: OpenAPI contract check
+        run: pnpm run openapi:check
+
+      - name: TypeScript type check
+        run: pnpm run typecheck:ts
+
   test-python:
     name: Python Tests (${{ matrix.os }})
     runs-on: ${{ matrix.runner }}

.gitignore

@@ -24,7 +24,11 @@ backend/**/__pycache__/
 .eggs/
 
 # Claude Code
-.claude/
+.claude/*
+!.claude/
+!.claude/skills/
+!.claude/skills/video-editor-development/
+!.claude/skills/video-editor-development/SKILL.md
 
 # IDE
 .idea/
@@ -65,6 +69,13 @@ models/*
 !models/README.md
 backend/models/
 
+# Profiling scripts and results
+backend/scripts/
+backend/profile_results/
+
+# Temporary working directory
+backend/tmp/
+
 # Generated outputs (videos and images)
 backend/outputs/*.mp4
 backend/outputs/*.png

AGENTS.md

@@ -34,7 +34,7 @@ PRs must pass: `pnpm typecheck` + `pnpm backend:test` + frontend Vite build.
 
 - **Path alias**: `@/*` maps to `frontend/*`
 - **State management**: React contexts only (`ProjectContext`, `AppSettingsContext`, `KeyboardShortcutsContext`) — no Redux/Zustand
-- **Routing**: View-based via `ProjectContext` with views: `home`, `project`, `playground`
+- **Routing**: View-based via `ProjectContext` with views: `home`, `project`
 - **IPC bridge**: All Electron communication through `window.electronAPI` (defined in `electron/preload.ts`)
 - **Backend calls**: Always use `backendFetch` from `frontend/lib/backend.ts` for app backend HTTP requests (it attaches auth/session details). Do not call `fetch` directly for backend endpoints.
 - **Styling**: Tailwind with custom semantic color tokens via CSS variables; utilities from `class-variance-authority` + `clsx` + `tailwind-merge`

README.md

@@ -30,10 +30,10 @@ LTX Desktop is an open-source desktop app for generating videos with LTX models
 
 | Platform / hardware | Generation mode | Notes |
 | --- | --- | --- |
-| Windows + CUDA GPU with **≥32GB VRAM** | Local generation | Downloads model weights locally |
-| Windows (no CUDA, <32GB VRAM, or unknown VRAM) | API-only | **LTX API key required** |
-| Linux + CUDA GPU with **≥32GB VRAM** | Local generation | Downloads model weights locally |
-| Linux (no CUDA, <32GB VRAM, or unknown VRAM) | API-only | **LTX API key required** |
+| Windows + CUDA GPU with **≥16GB VRAM** | Local generation | Downloads model weights locally |
+| Windows (no CUDA, <16GB VRAM, or unknown VRAM) | API-only | **LTX API key required** |
+| Linux + CUDA GPU with **≥16GB VRAM** | Local generation | Downloads model weights locally |
+| Linux (no CUDA, <16GB VRAM, or unknown VRAM) | API-only | **LTX API key required** |
 | macOS (Apple Silicon builds) | API-only | **LTX API key required** |
 
 In API-only mode, available resolutions/durations may be limited to what the API supports.
@@ -43,14 +43,14 @@ In API-only mode, available resolutions/durations may be limited to what the API
 ### Windows (local generation)
 
 - Windows 10/11 (x64)
-- NVIDIA GPU with CUDA support and **≥32GB VRAM** (more is better)
+- NVIDIA GPU with CUDA support and **≥16GB VRAM** (more is better)
 - 16GB+ RAM (32GB recommended)
 - **160GB+ free disk space** (for model weights, Python environment, and outputs)
 
 ### Linux (local generation)
 
 - Ubuntu 22.04+ or similar distro (x64 or arm64)
-- NVIDIA GPU with CUDA support and **≥32GB VRAM** (more is better)
+- NVIDIA GPU with CUDA support and **≥16GB VRAM** (more is better)
 - NVIDIA driver installed (PyTorch bundles the CUDA runtime)
 - 16GB+ RAM (32GB recommended)
 - Plenty of free disk space for model weights and outputs

backend/_routes/models.py

@@ -13,6 +13,8 @@
     ModelInfo,
     ModelsStatusResponse,
     RequiredModelsResponse,
+    TextEncoderAlreadyDownloadedResponse,
+    TextEncoderDownloadStartedResponse,
     TextEncoderDownloadResponse,
 )
 from _routes._errors import HTTPError
@@ -81,10 +83,10 @@ def route_text_encoder_download(handler: AppHandler = Depends(get_state_service)
 
     files = handler.models.refresh_available_files()
     if files["text_encoder"] is not None:
-        return TextEncoderDownloadResponse(status="already_downloaded", message="Text encoder already downloaded")
+        return TextEncoderAlreadyDownloadedResponse(status="already_downloaded", message="Text encoder already downloaded")
 
     session_id = handler.downloads.start_text_encoder_download()
     if session_id:
-        return TextEncoderDownloadResponse(status="started", message="Text encoder download started", sessionId=session_id)
+        return TextEncoderDownloadStartedResponse(status="started", message="Text encoder download started", sessionId=session_id)
 
     raise HTTPError(400, "Failed to start download")