VoiceKit MCP

Tools

• voicekit_analyze_acoustics Analyze acoustic features of audio. Extracts pitch, energy, rhythm, tempo, and spectral characteristics. Useful for understanding voice expressiveness and characteristics. Returns: pitch, energy, rhythm, tempo, spectral information
• voicekit_extract_embedding Extract voice embedding using Wav2Vec2. Returns a 768-dimensional vector representing voice characteristics. Useful for voice comparison, speaker identification, etc. Returns: embedding (768-dim list), model, dim
• voicekit_grade_voice Comprehensively compare and analyze user voice with reference voice. Evaluates with 5 metrics: - pronunciation: Pronunciation accuracy (STT-based) - tone: Voice timbre similarity (Wav2Vec2 embedding) - pitch: Pitch matching - rhythm: Rhythm sense - energy: Energy expressiveness. Returns: overall_score, metrics, weak_points, strong_points, feedback
• voicekit_isolate_voice Remove background music (BGM) and extract voice only. Uses ElevenLabs Voice Isolator to remove music, noise, etc. Useful for memes, songs, and other audio with background sounds. Returns: isolated_audio_base64, metadata (bgm_detected, sizes, duration)
• voicekit_match_voice Compare similarity between two voices. Extracts Wav2Vec2 embeddings and calculates cosine similarity. Useful for checking if the same person spoke with similar tone. Returns: similarity (0-1), tone_score (0-100)
• voicekit_transcribe_audio Convert audio to text (Speech-to-Text). Uses ElevenLabs Scribe v1 model for high-quality speech recognition. Supports various languages. Returns: text, language, model

No tool matches.

As the server's own listing gives them. We never call a tool.