Create a spoken dialogue with AI
Make a source-grounded host-and-expert dialogue and listen through an interactive browser speech player.
Create a spoken dialogue with AI when you want to hear an existing article or set of notes as a conversation. The podcast skill reads the source, groups its information and writes questions for a host with factual answers from an expert. A topic alone is insufficient: it requests source material rather than inventing details to sustain the exchange.
The deliverable is a React player that speaks the dialogue using voices available in the browser. You can pause, resume, stop or select a transcript line to continue there. Voice quality and language coverage depend on the device and browser. Some voices use online services, so playback is not necessarily offline. The result is a playable interface rather than an exported recording.
What it does
- Build a host-and-expert conversation from supplied prose
- Play an already formatted dialogue
- Organise exchanges by theme or chronology
- Prepare numbers, dates and abbreviations for speech
- Check the conversation against its factual source
- Return to a passage through its transcript line
Install and use
- Use Claude Artifacts or a compatible React renderer and a browser with speech synthesis. Supply authorised source material; host accounts are separate, and the skill needs no dedicated speech API key.
- Ask: “Make a host-and-expert conversation from this article, keeping only its facts.”
- Review names, figures and claims against the source before playback.
- Open the player, test pronunciation and request spoken-wording corrections where needed.
FAQ
Source
Made by sgasser. Original skill: podcast, published on GitHub under the MIT licence.
View on GitHub