LettuceAI/app desktop-dev-154-1-6a28ef5

Desktop Dev Build #154.1 (6a28ef5)

on GitHub

Automated desktop build artifacts for commit 6a28ef536f3f7fea22e5d653ba2f512fc434ac96.
Workflow run: https://github.com/LettuceAI/app/actions/runs/23917485227

Changes since previous Desktop build desktop-dev-151-1-a85d2e8

Compare: desktop-dev-151-1-a85d2e8...6a28ef5

• 6a28ef5 fix(windows): use DXGI video memory out param
• 7eb9e51 fix(windows): repair DXGI adapter checks
• d808a70 fix: fix typescript check error
• 93dedcc feat: add chat settings drawer and redesign session advanced settings
• becf3d2 refactor: migrate discovery pages to TopNav, add window controls to chat sub-pages
• 757b7b9 fix: add window controls and drag regions to all non-TopNav pages
• ac29e13 feat: custom window decorations with frameless titlebar
• 3170af1 refactor: redesign model editor layout for better horizontal space usage
• 6b9ccb2 refactor: replace embedded GGUF template modal with inline collapsible panel
• 31b1115 fix: expose llama.cpp reasoning settings
• a14a6de fix: share llama backend across runtime and context info
• 5491b01 feat: add local rp default prompt template
• 5c13103 fix: support additional thinking tag variants
• 8ca7834 fix: normalize thinking tags across api and local responses
• cd2c897 fix(llama-cpp): harden local tool-calling flow and structured-output failures
• ab00be4 fix: stabilize llama tool calling
• c16ed7d feat: add lorebook keyword detection modes
• ceab247 fix: clamp windows vulkan vram estimates
• 0776b51 fix: correct llama cpu fallback sizing
• 1ada467 fix: prevent settings reset after llama cpu fallback
• 0fa482e Merge pull request #40 from GentleBurr/feature/prompt-caching
• b7d8807 fix: track cache-aware usage and pricing
• 3f83920 fix: complete prompt caching support
• d02f523 refactor: remove legacy personas page
• 336a226 refactor: route persona flows through library
• 9ae7c66 fix: expand stacked sonner toasts
• 0294c62 fix: reuse llama backend for context info
• 045e302 feat: unload local chat model during dynamic memory
• 394af41 chore: fix memory state test error
• 46cdf36 feat: add llama model load progress toast
• 3db3c17 feat(memory): add progress bar to dynamic memory cycle
• 102d979 feat(chat): add prompt cache TTL and sticky routing
• 8156954 feat(models): add runability score to edit model page for llama.cpp models
• 5865d54 feat(ios): add onnxruntime installer workflow
• 459bfee feat(ollama): add native ollama integration
• fba299a Update README.md
• 833b133 fix(backup): preserve dynamic memory settings on restore
• 010c4a3 Update Cargo.lock