FluidVoice 1.5.1-beta.3

New Features

AI Thinking Tokens

• Added support for displaying AI "thinking tokens" (model reasoning)
• Improved parsing for thinking token blocks across different LLM providers
• Enhanced LLMClient to properly handle and separate thinking content from response content

Debug Logging Toggle

• Added a toggle in Settings to enable/disable debug logs
• Debug logger now respects the toggle setting for cleaner console output

Transcription History Toggle

• Added ability to enable/disable saving transcription history

Mode Shortcuts

• Added enable/disable toggle for Command Mode and Write Mode

Bug Fixes

Transcription Stability

• Fixed race condition checks during longer transcription sessions on some devices

Terminal Service

• Fixed JSON fallback escaping in TerminalService for more reliable command execution

Input Device Picker

• Fixed issue where the audio input device picker could sometimes appear empty