VoiceDeck MCP: Create AI Voice Presentations from Claude Code

Matthias Heim · 2026-05-13

How I wired a VAPI voice agent into a Model Context Protocol server so Claude Code can transcribe, summarise and act on spoken input without leaving the terminal.

What if you could turn any PDF into an AI-guided voice presentation with a single prompt? The VoiceDeck MCP server makes this possible directly from Claude Code, Cursor, or any MCP-compatible tool.

What is VoiceDeck?

VoiceDeck is an AI-powered presentation platform that transforms static documents into interactive, voice-guided experiences. Instead of reading through slides, your audience can have a conversation with an AI tutor that explains the content, answers questions, and adapts to their learning pace.

Upload any PDF or presentation document

AI analyzes content and generates teaching prompts

Recipients interact via natural voice conversation

Track engagement and comprehension metrics

Multiple AI voices to choose from (Gemini and OpenAI)

Why Use MCP for Voice Presentations?

The Model Context Protocol (MCP) allows AI assistants like Claude to interact with external services. With the VoiceDeck MCP server, you can create voice presentations without leaving your development environment.

Create presentations using natural language prompts

Integrate into CI/CD pipelines for automated content creation

No need to learn a new interface or API

Batch process multiple documents programmatically

Ask Claude to create a voice deck, and it just works.

How It Works

Configure MCP

Add the VoiceDeck MCP server to your Claude Code or Cursor configuration.

Prompt Claude

Use natural language to describe what voice deck you want to create.

Share the Link

Get a shareable link to your interactive voice presentation.

Available MCP Tools

create_voice_deck

Create a new voice deck with PDF upload, custom AI voices, and personalized teaching prompts.

get_voice_deck

Retrieve details of an existing voice deck including share URL and processing status.

Key Parameters

The create_voice_deck tool accepts these main parameters:

title - The name of your voice deck

description - What the deck covers

goal - The learning objective

voice_provider - Choose between Gemini or OpenAI

Gemini voices: Puck, Charon, Kore, Fenrir, Aoede

OpenAI voices: alloy, echo, fable, onyx, nova, shimmer

pdfUrl or pdfBase64 - Your presentation content

Example: Creating a Voice Deck

Here's a natural conversation with Claude to create a voice deck:

Create a voice deck from my quarterly report PDF. Use a friendly tone for the sales team. Title it 'Q4 2024 Results Overview'.

Creating your voice deck... ✓ Title: Q4 2024 Results Overview ✓ Voice: Puck (friendly Gemini voice) ✓ Processing PDF... Share link: voicedeck.ai/shared/abc123

Real-World Use Cases

Automated Onboarding

Every time your employee handbook is updated, automatically generate a voice deck. New hires can ask questions about policies instead of reading 50 pages.

Sales Enablement

Create voice decks for proposals. Prospects can have the deck explain itself. No more scheduling demo calls for every lead.

Documentation That Talks

Generate voice decks from API docs. Developers can ask 'how do I authenticate?' instead of searching through pages.

Quick Setup Guide

1. Get an API Key

Sign up at voicedeck.ai and get your API key from settings.

2. Configure Claude Code

Add the VoiceDeck MCP server to your configuration file.

3. Start Creating

Ask Claude to create a voice deck with your content.

Add to your .mcp.json or Claude Code settings:

Transform Your Content Delivery

Stop creating static presentations. Start creating interactive voice experiences that engage your audience and answer their questions in real-time.

How we work · Work · Insights