Add Local AI Support, fix minor bugs

• Added new Local provider for self-hosted, OpenAI-compatible LLM servers (vLLM, llama.cpp, Ollama, LM Studio, etc..)
• Configure via a custom Base URL - API key is optional for local servers
• Max tokens now defaults to a safe 32K for non-DeepSeek providers to prevent errors; DeepSeek keeps its 384K default

v0.1.1 BETA

v0.1.1

• Fix: console windows no longer flash on Windows (child processes spawn hidden)
• Add: LucidQuery provider with agi-01-swift and agi-01-frontier models
• Use actual model names instead of hardcoded values

CodeDone v0.1.0 - Windows x64

v0.1.0 BETA