---
title: Multi-Speaker Audiobook Generator
emoji: "\U0001F3AD"
colorFrom: purple
colorTo: indigo
sdk: gradio
sdk_version: 5.25.0
app_file: app.py
pinned: false
license: mit
---

# Multi-Speaker Audiobook Generator

Upload a story and get an audiobook where each character has a unique voice.

## Features

- Auto-detect characters from dialogue or define them manually
- Up to 8 unique voices per story
- 17 languages across two AI engines
- Emotional audio tags for expressive narration
- PDF, DOCX, TXT input support

## Setup

Add these secrets in Settings:

| Secret | Required | Used for |
|--------|----------|----------|
| DASHSCOPE_API_KEY | Yes | Character detection, translation, premium voices |
| ELEVENLABS_API_KEY | For English accents | US/UK/AU English + African languages |

## How It Works

1. Upload or paste your story
2. Click "Detect Characters" - AI finds all speakers
3. Assign voices to each character
4. Generate - each character gets their own voice