MCP FishAudio Server

# Changelog All notable changes to this project will be documented in this file. The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). ## [0.6.1] - 2025-01-04 ### Fixed - Fixed MCP JSON communication issues by replacing console output with logger utility - Logger only outputs when DEBUG environment variable is set - Prevents stdout/stderr messages from interfering with MCP JSON-RPC protocol ## [0.6.0] - 2025-01-03 ### Added - Multiple voice reference management system - New tool: `fish_audio_list_references` to list configured voices - Voice selection by name or tag in addition to ID - Support for configuring multiple voices with metadata - Added FISH_REFERENCES and FISH_DEFAULT_REFERENCE environment variables - Enhanced voice selection with priority: ID > Name > Tag > Default ### Changed - References can now be configured as JSON array or individual environment variables - Improved voice selection logic with metadata support ## [0.5.4] - 2025-01-03 ### Fixed - Fixed zod version compatibility by pinning to exact version 3.23.8 - Resolved version conflict between MCP SDK and Fish Audio SDK dependencies - Verified dev and build modes work correctly locally ## [0.5.3] - 2025-01-03 ### Fixed - Added missing `zod` dependency to fix "Cannot find module" error - Fixed module resolution issues when running via npx ## [0.5.2] - 2025-01-03 ### Fixed - Fixed audio playback not working when FISH_STREAMING=true and FISH_AUTO_PLAY=true - Fixed tilde (~) expansion in AUDIO_OUTPUT_DIR environment variable - Separated HTTP streaming from WebSocket streaming to prevent connection issues - WebSocket streaming now requires explicit enabling via tool parameters ### Changed - FISH_STREAMING now only controls HTTP streaming (not WebSocket) - WebSocket streaming defaults to false to improve stability ## [0.5.1] - 2025-01-03 ### Changed - Improved README documentation formatting - Changed "Required" column in environment variables table to use text instead of emojis - Updated references from "Claude Desktop" to "MCP Settings" for broader compatibility - Added AUDIO_OUTPUT_DIR to the configuration example - Updated FISH_STREAMING description to clarify it controls both HTTP and WebSocket streaming - Updated FISH_AUTO_PLAY description to clarify it controls both auto-play and real-time playback ## [0.5.0] - 2025-01-03 ### Breaking Changes - Removed `FISH_WEBSOCKET_STREAMING` environment variable - Removed `FISH_REALTIME_PLAY` environment variable - WebSocket streaming is now controlled by `FISH_STREAMING` - Real-time playback is now controlled by `FISH_AUTO_PLAY` ### Changed - Simplified environment variable configuration - `websocket_streaming` parameter now defaults to `FISH_STREAMING` value - `realtime_play` parameter now defaults to `FISH_AUTO_PLAY` value - Cleaner configuration with unified controls ### Migration Guide - Remove `FISH_WEBSOCKET_STREAMING` from your config (use `FISH_STREAMING` instead) - Remove `FISH_REALTIME_PLAY` from your config (use `FISH_AUTO_PLAY` instead) ## [0.4.1] - 2025-01-03 ### Added - Environment variable mapping for WebSocket streaming - `FISH_WEBSOCKET_STREAMING` defaults to `FISH_STREAMING` value - `FISH_REALTIME_PLAY` defaults to `FISH_AUTO_PLAY` value - Simplified configuration with intelligent defaults ### Changed - WebSocket streaming can now be enabled via environment variables - Real-time playback can now be enabled via environment variables - Improved configuration documentation ## [0.4.0] - 2025-01-03 ### Added - Integration with official Fish Audio SDK - Proper latency parameter support (normal/balanced) - Config caching for better performance ### Changed - Refactored entire codebase to use Fish Audio SDK - Improved WebSocket streaming stability - Better error handling with SDK error types - Simplified API calls using SDK methods ### Fixed - Auto-play functionality now works correctly - Connection stability issues resolved - Memory efficiency improvements ### Removed - Custom API implementation (replaced by SDK) - Direct axios calls (handled by SDK) ## [0.3.0] - 2025-01-03 ### Added - WebSocket streaming support for real-time TTS via `wss://api.fish.audio/v1/tts/live` - Real-time audio playback during WebSocket streaming with `realtime_play` parameter - `websocket_streaming` parameter to enable WebSocket mode - `FishAudioWebSocketService` for handling WebSocket connections - `RealTimeAudioPlayer` utility for immediate audio output - Support for both HTTP and WebSocket streaming modes - MessagePack encoding/decoding for WebSocket communication ### Changed - TTS tool now supports three modes: standard, HTTP streaming, and WebSocket streaming - Response includes `streaming_mode` and `total_bytes` information - Updated documentation with WebSocket examples ### Technical Details - WebSocket uses MessagePack for message encoding - Real-time player supports multiple platforms (ffplay, mpv, afplay) - Text is chunked for optimal WebSocket streaming performance ## [0.2.0] - 2025-01-03 ### Added - Automatic audio playback feature with `auto_play` parameter - Cross-platform audio playback support (macOS, Windows, Linux) - FISH_AUTO_PLAY environment variable for default playback behavior - Audio player utility for handling platform-specific playback ### Changed - TTS tool now returns `played` status in response - Updated documentation with auto-play feature ### Technical Notes - Streaming mode uses HTTP streaming API (not WebSocket) - Audio playback is handled through native OS commands ## [0.1.2] - 2025-01-03 ### Changed - Changed npm package name to @alanse/fish-audio-mcp-server for scoped publishing - Updated documentation to reflect new package name ## [0.1.1] - 2025-01-03 ### Fixed - Fixed directory creation error when running via npx - Changed default audio output directory to `~/.fish-audio-mcp/audio_output` for better permissions - Added error handling for directory creation failures ### Changed - Audio output directory now defaults to user's home directory instead of relative path - Improved error messages for directory creation issues ## [0.1.0] - 2025-01-03 ### Added - Initial release - Basic TTS functionality with Fish Audio API integration - Streaming support for real-time audio generation - Environment variable based configuration - Support for multiple audio formats (MP3, WAV, PCM, Opus) - MCP tools for text-to-speech generation - Custom voice support via reference IDs