Webcast

Name: Webcast
Rating: 92
Author: damir5

Convert web articles to MP3 audio, markdown, or DOCX using local TTS on Apple Silicon.

Commands

bash

# Extract article text from a URL
webcast extract <url>                          # stdout (plain text)
webcast extract <url> -o article.txt           # to file
webcast extract <url> --format md              # as markdown
webcast extract <url> --format json            # as JSON

# Convert text to speech (Chatterbox by default)
webcast tts article.txt -o output.mp3          # from file
echo "Hello world" | webcast tts -o hello.mp3  # from stdin
webcast tts article.txt --ref-audio voice.wav  # clone a voice
webcast tts article.txt --model kokoro --voice af_bella  # use Kokoro

# One-step: URL to MP3/markdown/DOCX
webcast convert <url>                          # MP3 (Chatterbox), auto-named in ./output/
webcast convert <url> --model kokoro           # MP3 with Kokoro
webcast convert <url> --format md              # rich markdown (links, images, tables)
webcast convert <url> --format docx            # DOCX via pandoc
webcast convert <url> -o episode.mp3           # custom output path

# Pipe: extract then speak
webcast extract <url> | webcast tts -o out.mp3

TTS Models

Chatterbox (default)

•Expressive, natural-sounding speech
•Voice cloning via --ref-audio <wav> (5+ seconds of reference audio)
•Uses default voice when no reference audio provided

Kokoro

•Fast, lightweight (82M params)
•Named voice presets via --voice
•Speed control via --speed

Kokoro Voice Presets

American English:

•af_heart (default), af_bella, af_nova — female
•am_adam, am_echo — male

British English:

•bf_alice, bf_emma — female
•bm_daniel, bm_george — male

Options

Flag	Default	Description
`--model`	`chatterbox`	TTS model: chatterbox, kokoro
`--ref-audio`	—	Reference audio for Chatterbox voice cloning
`--voice`	`af_heart`	Kokoro voice preset
`--speed`	`1.0`	Kokoro speed multiplier (0.5–2.0)
`--output-dir`	`./output`	Default output directory
`-o, --output`	auto	Output file path
`--format` (extract)	`txt`	Extract format: txt, json, md
`--format` (convert)	`mp3`	Convert format: mp3, md, docx

Agent Usage

To convert a blog post (uses Chatterbox by default):

code

webcast convert https://example.com/blog-post

To use Kokoro with a specific voice:

code

webcast convert https://example.com/blog-post --model kokoro --voice af_heart

To save as markdown or DOCX:

code

webcast convert https://example.com/blog-post --format md
webcast convert https://example.com/blog-post --format docx

Requirements

•macOS with Apple Silicon
•ffmpeg installed (brew install ffmpeg)
•pandoc installed (brew install pandoc) — for DOCX output
•First run downloads Chatterbox Turbo model (~1GB) or Kokoro model (~170MB)