Support Audio Modality

[evalstate]

Pre-submission Checklist

• I have verified this would not be more appropriate as a feature request in a specific repository
• I have searched existing discussions to avoid duplicates

Your Idea

MCP currently supports Images as a first-class content type alongside text in Prompts/Sampling. Although Audio can be transmitted as an embedded resource I think it would be better to treat it as a first-class modality rather than a generic resource type.

The current reference Client implementation (Claude Desktop) and associated models do not currently support Audio - although I think the MCP protocol should.

I'd propose something like the below:

export interface AudioContent extends Annotated {
  type: "audio";
  /**
   * The base64-encoded audio data
   */
  data: string;
  /**
   * The MIME type of the audio. Different providers may support different audio formats.
   */
  mimeType: string;
}

There's also potential for adding optional metadata for Image/Audio types similar to below to help Client integration.

  /**
   * Optional metadata about the audio like duration, sample rate, etc.
   */
  metadata?: {
    durationMs?: number;
    sampleRate?: number;
    channels?: number;
  };
}

I'd be happy to create PR if in agreement. PR submitted #93

Scope

• Protocol Specification
• SDK Features
• Documentation
• Developer Experience
• Other