v1.3.0 — Local LLM & Offline AI

What's New

Local LLM Support

• Download and run Qwen 2.5-7B (or any GGUF model) entirely offline
• Zero-Dependency Download: Pure Python stdlib + shell/curl fallback — no pip install needed
• Smart GGUF selection: auto-picks best single-file quantization (q3_k_m > q2_k > ...)
• Built-in health monitoring, warmup progress, and crash recovery

Code-Enforced Group Chat Routing

• Group chat routing now deterministic (code decides who speaks to whom)
• LLM only generates content — identical conversation flow across cloud and local models
• User speaks → pets discuss (2-3 rounds) → wrap up to user

Improvements

• pet-llm-server.py: Shell/Python polyglot — works with /bin/sh even without Python
• pet-group-chat.py: Local LLM support with short prompt optimized for 7B models
• pet-auto-reply.py: Auto-detect local vs cloud provider
• Settings UI: New Local LLM tab with model download, server start/stop, health status

Technical

• Supports macOS 12+ (macOS 13+ ships with built-in Python 3)
• Bundled llama-server for Apple Silicon acceleration (Metal GPU)
• ~3.5GB default model download (Qwen 2.5-7B GGUF q3_k_m)

v1.2 — Group Chat: Dual-Pet Autonomous Discussions

🆕 Group Chat (v1.2)

Two pets, one stage. Duck and Capybara appear simultaneously and engage in autonomous multi-turn discussions.

What's New

• Dual-Pet Group Chat: Open from menu bar, configure both pets' personalities, start a topic
• Autonomous Pet-to-Pet Dialogue: User speaks → pets discuss among themselves (2-3 rounds) → then respond
• Color-Coded Bubbles: Duck (warm cream 🟡), Capybara (ice blue 🔵), User (white ⬜)
• Hover Glow Effects: Mouse over any bubble for a subtle highlight
• Collapsible Bubble Stacks: Auto-fold with "▼ N more..." expander
• Rich Conversation Log: Bold speaker names, proper spacing
• AI Session Summary: Stop the session to receive an AI-generated discussion summary
• Round Counting: Fixed bug where Capybara would skip replying in round 2+
• Clean Stop: All UI residues properly cleaned when ending a session

Bug Fixes

• Capybara now always participates in round 2+ discussions
• User input window color now distinctly white (not matching Duck's cream)
• Stop button properly clears all bubble/fold/input windows

Files

• duck-pet.swift — Main app (+675 lines for Group Chat)
• pet-group-chat.py — AI orchestration engine (new)
• README.md — Updated with Group Chat section and screenshot

Download

Grab DesktopDuck-v1.2.app.zip below, unzip, and run!

Requirements

• macOS 12+ (native AppKit)
• Python 3 + Pillow
• MiniMax API key

Desktop Duck v1.1

What's New

Journal

• Rewritten as iterative document model — sessions enrich a growing markdown document via AI summarization
• Template selector, Refresh/+Session/Continue buttons, live markdown preview

Spritesheet Editor

• Draggable grid lines for precise frame adjustment
• GIF animation preview after conversion
• Auto-trim content: frames crop to actual character bounds, no more tiny/missing pets
• Dynamic aspect ratio selection based on grid dimensions

Pet Switching

• Duck 🦆 / Capybara 🦫 toggle in Preferences (Appearance section)
• Original duck-idle.gif backed up before custom apply, restored on switch

Bug Fixes

• Uploaded spritesheet now shows in preview immediately
• Show Grid checkbox works on first open
• Preferences scrolls to top on open (no more scrolled-to-bottom)
• AI Model label layout fixed (labels above inputs)
• Slider descriptions now above sliders (Width, Height, etc.)

Python Scripts

• pet-generate-character.py: dynamic aspect ratio, background removal, auto-trim
• pet-convert-spritesheet.py: custom row/col splits, auto-trim
• pet-journal-summary.py: iterative document AI summarization

Config

• New petType field (duck/capybara/custom)
• Save AI Config button in Preferences

Full changelog: see commits since initial release.

DesktopDuck v1.0 — Initial Release

DesktopDuck v1.0 🦆

Initial release of the AI-powered desktop pet duck for macOS.

Features

• AI Chat — Double-click to chat via MiniMax or any OpenAI-compatible API
• Journal System — 5 mindfulness templates with AI-generated summaries
• AI Character Generation — Describe a character, generate a custom pet image
• Fully Configurable — Size, bubbles, timeout, context window, window level
• Window Levels — Always on Top / Normal / Always at Bottom
• Bubble Stack — Floating message bubbles with auto-expiry

Requirements

• macOS 13+ (Ventura or later)
• Python 3.7+ with Pillow (pip3 install Pillow)
• MiniMax API key (or any OpenAI-compatible API)

Installation

1. Download DesktopDuck-v1.0.zip
2. Unzip and move DesktopDuck.app to Applications
3. Right-click → Open (first launch only)
4. Open Preferences (⌘,) and enter your API key
5. Double-click the duck to start chatting!

Or build from source — see README.