dz-podcast
Generate and publish a dual-host daily podcast. Fetches news, generates a conversational script between two hosts, synthesizes audio via Fish Audio or Edge TTS, publishes to S3 with RSS feed for Apple Podcasts, Spotify, etc. Fully automated with cron support.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/dz1922/dz-podcastDual-Host Daily Podcast Generator
Automated daily podcast with two AI hosts. Generates text brief + dual-voice audio, publishes to RSS, delivers via messaging.
Concept
- Format: Two hosts — one explains/analyzes, the other asks and transitions
- Duration: Configurable, default ~7 minutes
- Style: Casual, opinionated, conversational — like two friends chatting about the news
- Topics: Customizable (default: AI/Tech, Stocks, Macro, Crypto)
Architecture
Fetch News → Text Brief → Dual-Voice Script → TTS Audio → S3 Upload → RSS Update → Deliver
Configuration
Set these in your environment:
| Variable | Description |
|---|---|
S3_BUCKET | S3 bucket name |
PODCAST_DOMAIN | Custom domain or S3 URL |
FISH_API_KEY | Fish Audio API key (https://fish.audio) |
FISH_VOICE_A | Fish Audio voice ID for Host A |
FISH_VOICE_B | Fish Audio voice ID for Host B |
Step 1: Fetch News
Use web_fetch to scrape sources in parallel. Default sources:
https://news.ycombinator.com/— Techhttps://www.coindesk.com/— Cryptohttps://techcrunch.com/category/artificial-intelligence/— AIhttps://finance.yahoo.com/— Markets
Customize sources to match your podcast topic.
Step 2: Generate Text Brief
Organize news into sections with emoji headers:
☀️ Daily Brief | Mar 3, 2026
━━━━━━━━━━━━━━━━━━
🤖 Tech / AI
① Headline
→ One-line take
━━━━━━━━━━━━━━━━━━
📈 Markets
① Headline
→ One-line take
━━━━━━━━━━━━━━━━━━
🎯 Key Takeaway
Summary paragraph
Step 3: Generate Dual-Voice Script
Rewrite the brief as a dialogue. Prefix each line with speaker tag:
HostA: Welcome to today's episode...
HostB: Some big stories today...
HostA: Right, let's start with...
Guidelines:
- Host A: Explains and analyzes, knowledgeable but casual
- Host B: Asks, transitions, reacts
- Substantial turns, not one-liners
- Include analysis and discussion, not just headlines
- End with a lighter topic + sign-off
Step 4: Generate Audio
Fish Audio (recommended — natural, multi-voice):
python3 scripts/fish_dual_tts.py <script.txt> <output.mp3>
Parses speaker tags, sends each segment to Fish Audio, concatenates into final MP3.
Edge TTS (free fallback, single voice):
edge-tts --voice en-US-GuyNeural --rate "+5%" --file script.txt --write-media output.mp3
Step 5: Publish
bash scripts/generate_episode.sh <date> <EP-number> <title> <description> <mp3-file>
What it does:
- Upload MP3 to S3
- Get actual duration via ffprobe
- Insert
<item>into RSS feed (newest first) - Update
<lastBuildDate>
Step 6: Deliver
Send text brief + audio via your preferred channel (Telegram, Discord, Slack, etc.)
RSS Feed
See references/rss-format.md for XML template.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-dz1922-dz-podcast": {
"enabled": true,
"auto_update": true
}
}
}