MCP Server Whisper

"""Response models for MCP tool outputs.""" from typing import Any, Optional from pydantic import BaseModel, Field class AudioProcessingResult(BaseModel): """Result from audio processing operations (convert, compress).""" output_file: str = Field(description="Name of the processed audio file") class TranscriptionResult(BaseModel): """Result from transcription operations. Includes all fields that OpenAI's transcription API might return. """ text: str = Field(description="The transcribed text") duration: Optional[float] = Field(default=None, description="Duration of the audio in seconds") language: Optional[str] = Field(default=None, description="Detected language of the audio") segments: Optional[list[dict[str, Any]]] = Field(default=None, description="Timestamped segments") words: Optional[list[dict[str, Any]]] = Field(default=None, description="Word-level timestamps") usage: Optional[dict[str, Any]] = Field(default=None, description="Token usage information") logprobs: Optional[Any] = Field(default=None, description="Log probabilities if requested") model_config = {"arbitrary_types_allowed": True} class ChatResult(BaseModel): """Result from chat_with_audio operation.""" text: str = Field(description="The response text from the audio chat") class TTSResult(BaseModel): """Result from text-to-speech operations.""" output_file: str = Field(description="Name of the generated audio file")