Read Any Web Page with OpenClaw
CastReader is the only OpenClaw skill that extracts content from any URL and reads it aloud with natural AI voices. Send a link via Telegram — get an MP3. No other ClawHub skill can do this.
See It in Action
Send a URL, get paragraph-by-paragraph audio + text back
Get Started in 3 Steps
From install to audio in under a minute
Install
Search "castreader" on ClawHub and install the skill, or run clawhub install castreader in your terminal.
Send a URL
In Telegram, Discord, or Slack — send any article URL to your OpenClaw agent.
Get Audio + Text
The agent extracts content and sends back each paragraph as a separate message: the paragraph text as a caption, plus its MP3 audio. You can read along while listening — or just play the combined full.mp3.
Commands
Three commands for different workflows
extract <url>
Extract article content as structured JSON — useful for summarization, research, or piping into other skills.
extract https://example.com/article
Returns JSON with title, paragraphs[], and detected language.
generate-audio <url>
Extract content and generate a separate MP3 for each paragraph, plus a combined full.mp3. Returns a JSON manifest mapping each paragraph's text to its audio file — perfect for sending as individual messages on Telegram/Discord/Slack.
generate-audio https://example.com/article
Per-paragraph: 001.mp3 + text, 002.mp3 + text, ... Combined: full.mp3
read-aloud <url>
Open the page in a browser and read aloud with real-time paragraph highlighting.
read-aloud https://example.com/article
Triggers in-browser reading with paragraph-level highlighting and auto-scroll. Requires CastReader Chrome extension.
What CastReader Does That No Other Skill Can
The only OpenClaw skill that goes from URL to audio
URL to Audio
Send a URL, Get Paragraph-by-Paragraph Audio
Send any URL to your OpenClaw agent via Telegram, Discord, or Slack. CastReader extracts the article text, converts each paragraph to speech with the Kokoro TTS model, and sends back each paragraph's text + MP3 individually — so you can read along while listening. Also generates a combined full.mp3 for offline playback.
Smart Extraction
Works Where Others Fail
CastReader has dedicated extractors for 15+ platforms that break generic tools: Kindle Cloud Reader (font decoding), WeRead (canvas rendering), Notion, Google Docs, Medium, Substack, ChatGPT, Claude, arXiv, and more.
Browser Mode
In-Browser Reading with Highlighting
When used with the CastReader Chrome extension, OpenClaw can trigger in-browser reading with paragraph-level highlighting. Each paragraph lights up as it's read, with automatic scrolling.
CastReader vs Other TTS Skills
See how CastReader compares to other text-to-speech skills on ClawHub
| Skill | Web Extraction | URL to Audio | 40+ Languages | Highlighting | Free |
|---|---|---|---|---|---|
| CastReader | |||||
| kokoro-tts | |||||
| openai-tts | |||||
| mac-tts |
Frequently Asked Questions
Everything about using CastReader with OpenClaw
How do I install the CastReader OpenClaw skill?
Go to clawhub.ai/vinxu/castreader and click Install, or search for 'castreader' on ClawHub. You can also run 'clawhub install castreader' in your terminal. The skill will be available to your OpenClaw agent immediately.
What is OpenClaw / ClawHub?
OpenClaw is an open protocol for AI agent skills — it lets agents use tools like CastReader through a standard interface. ClawHub (clawhub.ai) is the marketplace where you discover, install, and manage OpenClaw skills. Think of it as npm for AI agents.
How is this different from kokoro-tts or openai-tts?
kokoro-tts and openai-tts only convert plain text strings to audio. CastReader extracts content from web pages — it handles complex DOM structures, canvas-rendered text (WeRead), scrambled fonts (Kindle), and 15+ platform-specific extraction methods. Then it converts the extracted text to speech.
Which messaging platforms does it work with?
Any platform OpenClaw supports: Telegram, Discord, Slack, WhatsApp, and more. Send a URL, get an MP3 audio file back.
Does it need the Chrome extension?
For URL-to-audio conversion, no — the skill works standalone. For in-browser reading with paragraph highlighting, you need the CastReader Chrome extension installed in the OpenClaw-controlled browser.
Is it free?
Completely free. No API key needed, no usage limits, no account required.
What websites does it work on?
Any website. CastReader has dedicated extractors for Kindle Cloud Reader, WeRead, Notion, Google Docs, Medium, Substack, arXiv, Wikipedia, ChatGPT, Claude, Gemini, and more. For all other sites, it uses a smart visible-text-block algorithm.
What languages are supported?
40+ languages with natural AI voices powered by the Kokoro TTS model.
Can I change the voice or speed?
Yes. Set environment variables before running: CASTREADER_VOICE (default: af_heart) to choose a voice, and CASTREADER_SPEED (default: 1.5) to adjust speech rate. The Kokoro model supports many voice options across 40+ languages.
Can I use it to extract text without audio?
Yes. The extract command returns structured JSON with title, paragraphs, and detected language — useful for research, summarization, or piping into other skills. It runs entirely locally with no API calls.
Do extract and generate-audio need the Chrome extension?
No. Only the read-aloud command needs the Chrome extension (for in-browser highlighting). Extract and generate-audio are fully standalone — they use Puppeteer headless and work on any server without a browser UI.
Start Listening Now
Completely free. No signup. No limits. Install and start listening.