One MCP server. Text from your Hermes agent in, a published podcast episode out, in your favourite podcast app.
Or skip reading. This page, made into a podcast by Hermes itself.
No backend to wire up. No file hosting. No RSS to publish. The MCP server does it all.
Briefing, research summary, weekly digest. Whatever your Hermes agent already produces.
One tool call from inside your agent. The cast0 MCP server handles auth, the API, and polling for completion.
Auto-published to your RSS feed. Apple Podcasts, Spotify, Overcast, anywhere you already listen.
You set up your agent to research, summarize, brief, digest. It does the work. The output piles up in chat scrollback. You read the first few lines and move on.
Audio fixes this. Episodes queue up while you sleep. The morning briefing waits in your podcast app next to The Daily. The research summary plays on your commute. The weekly digest catches you on a walk.
Your agent talks when it has something to say. You listen when you have time.
From zero to a published episode in about two minutes.
Use the prompt or CLI at the top of this page.
Sign up at cast0.ai, name your podcast, pick a voice. Free, no card. The dashboard shows your API key (pk_...) and the RSS URL.
Store it as CAST0_API_KEY in ~/.hermes/.env (or run hermes config set CAST0_API_KEY pk_...). Hermes substitutes it into the MCP URL at connect time.
Try: "Brief me on AI news from this week and ship it as an episode."
Full walkthrough with all three setup paths (prompt, CLI, config.yaml) in the Connect Hermes docs. The free tier covers about 40 minutes of audio per month. Need more? Paste your own OpenAI TTS key on the settings page and episodes stop counting against quota.
hermes mcp add, or edit ~/.hermes/config.yaml directly.One MCP. Real RSS. Episodes in the podcast app you already use.