cast0 MCP server

MCP server for creating podcast episodes from text. Five tools: create episodes, check status, list episodes, and manage your podcast's name and voice. Connect any MCP-compatible AI client.

Get an API key

MCP server URL

https://api.cast0.ai/mcp

Streamable HTTP transport. Authenticate by appending your podcast API key (pk_...) to the URL as ?key=pk_.... Clients that support OAuth can instead add the bare URL and sign in with your cast0 account. Find your key in the dashboard.

Supported AI agents

Connect cast0 with any MCP-compatible client to create podcast episodes.

Claude Desktop

Connect via the JSON configuration file. Claude Desktop reaches remote HTTP servers through the mcp-remote bridge.

  1. Open Settings → Developer and click Edit Config:
    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • Windows: %APPDATA%\Claude\claude_desktop_config.json
  2. Replace or update the file contents with the JSON below.
  3. Restart Claude Desktop to load the new configuration.
claude_desktop_config.json
{
  "mcpServers": {
    "cast0": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://api.cast0.ai/mcp?key=pk_YOUR_KEY"
      ]
    }
  }
}

For more details, see Claude Desktop MCP.

Tools

create_episode(title, text, teaser?)

Generate a new episode from text. The optional teaser is a 1-2 sentence hook spoken as a cold open at the top of the episode, mixed with the intro music. Returns immediately with status queued. Auto-publishes to the podcast's RSS feed when ready. In clients that support MCP Apps, shows an inline player that tracks rendering automatically.

get_episode(id)

Fetch a single episode by id. Poll until status is done to get the audio URL.

list_episodes

List every episode for the podcast, newest first.

get_podcast

Get the podcast's settings and links: name, TTS model, current voice and the voices available for it, music setup, the RSS subscribe link, and the dashboard settings URL.

update_podcast(name?, voice?)

Rename the podcast or switch its TTS voice. Model, speed, and music stay in the dashboard; the tool hands out the settings URL instead.

Every tool also takes an optional podcast_id. On OAuth connections it selects between the account's podcasts (omit it to use the default). API key connections are always scoped to the key's podcast.