v1.1.0: Heuristic Voice Selection & Artifact Rendering

What's New in v1.1.0

Voice Selection Improvements

• Fixed Chrome selecting poor quality voices (Albert, Bahh, etc.)
• Replaced hardcoded voice lists with intelligent heuristic-based scoring
• Works automatically for all 250+ voices and all languages
• Quality scoring: Neural/Natural (500pts), Wavenet (400pts), Microsoft (250pts), Google (200pts)
• Added blacklist for novelty voices

Artifact Rendering

• Clearer instructions for artifact rendering
• Added Write tool to allowed-tools
• Simplified JSX file creation process

Features

• Multi-language support (EN, DE, FR, ES, IT, PT, NL, PL, RU, JA, ZH)
• Browser-based TTS with zero cost, no API keys
• Interactive podcast player with play/pause/resume controls
• Platform-aware voice selection (iOS, Android, Desktop Edge, Desktop Chrome)

Installation

1. Download podcast.skill from this release
2. Open Claude (Desktop, Web, or Mobile)
3. Go to Settings → Capabilities → "Install from file"
4. Select podcast.skill

Usage

Simply ask Claude to create a podcast:

• "Create a podcast about the history of artificial intelligence"
• "Create a podcast from this article: [paste article]"
• "Erstelle einen deutschen Podcast über Quantencomputer"

Tip: Use Microsoft Edge browser for best voice quality (250+ high-quality voices vs Chrome's 19).

Full Changelog

v1.1.0 (this release)

• Heuristic-based voice selection
• Blacklist for novelty voices
• Clearer artifact rendering instructions

v1.0.0 (2024-11-23)

• Initial release