podcast-generator

Generate a twin-host Chinese podcast from topic to HTML player.

2|1|Updated Mar 4, 2026

Search Tags:#automation #dialogue #podcast #tts #multi-host #html-player

Ready to run

npx skills add https://github.com/cncoder/oneclaw --skill podcast-generator-cncoder

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: podcast-generator
Source: https://github.com/cncoder/oneclaw/tree/main/skills/podcast-generator
Command: npx skills add https://github.com/cncoder/oneclaw --skill podcast-generator-cncoder

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires loqui_tts, mlx-audio, boto3, numpy, soundfile, torch, qwen_tts, pyyaml, pyloudnorm, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill automates end-to-end podcast production, turning topic ideas into research, scripts, TTS audio, and an HTML5 player, all locally to streamline creator workflows.

Core Features & Use Cases

Local, cross-platform twin-host podcast pipeline: macOS Apple Silicon and Linux x86_64 supported; deterministic research → script → TTS → HTML player; optional private CloudFront publishing.
On-demand resources and modular workflow: choose to generate research, write scripts, synthesize audio, produce HTML, and publish privately to CDN.
Safe, auditable publishing with private buckets and CloudFront access controls, including optional ASR checks and chunk caching for breakpoint resume.

Quick Start

Run the one_shot workflow with a topic to generate a complete podcast including script, TTS audio, and an HTML player.

Frequently Asked Questions about podcast-generator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate generating a Chinese podcast from a topic locally?▼

To automate Chinese podcast generation locally, provide a topic to trigger a deterministic workflow that performs research, scripts a twin-host dialogue, synthesizes TTS audio, and outputs a publish-ready HTML5 player.

Can I generate a multi-host podcast script and audio on macOS Apple Silicon?▼

Yes, multi-host podcast generation is supported on macOS Apple Silicon using the MLX backend, alongside Linux x86_64 using Torch, to locally synthesize twin-host dialogue and audio.

What's the best way to publish a locally generated podcast audio privately?▼

The best way to publish locally generated podcast audio privately is to upload the HTML player and audio files to private S3 buckets with CloudFront access controls for safe, auditable CDN distribution.

Do I need loqui_tts and AWS boto3 to run the podcast workflow?▼

Yes, loqui_tts is required for audio synthesis and boto3 is needed for AWS CloudFront utilities, alongside MLX or Torch backends, to execute the end-to-end podcast workflow.

How does the podcast pipeline handle breakpoint resume for audio synthesis?▼

The podcast pipeline handles breakpoint resume during audio synthesis by implementing chunk caching, allowing the deterministic workflow to pause and recover progress without restarting the TTS generation entirely.