Connect Claude, ChatGPT and other AI assistants (MCP)
Add the Jellypod MCP server to Claude, ChatGPT, Claude Code, Codex, Zapier or n8n, sign in, and see what the assistant can and cannot do.
The Jellypod MCP server lets an AI assistant create and manage content in your library. Your plan limits and credit rates still apply.
Connect
Add this remote server to your client:
URL: https://mcp.jellypod.com/mcp
Transport: Streamable HTTP
Auth: OAuthSign in to Jellypod when the client asks. The connection uses your current library and permissions. Clients that cannot complete OAuth send a Jellypod API key as an Authorization: Bearer YOUR_API_KEY header.
- OAuth sign-in: MCP Connector feature on Starter, Educator, Creator, Business and Enterprise
- API-key bearer authentication: API access (Creator, Business and Enterprise)
| Client | Setup |
|---|---|
| Claude (web and desktop) | Open Settings, then Connectors, click Add custom connector, paste the URL, then click Connect and approve access. |
| ChatGPT (web) | Turn on Developer mode under Advanced settings in Apps & Connectors, add a new app with the URL and OAuth authentication, then approve access. |
| Claude Code | Run claude mcp add --transport http jellypod https://mcp.jellypod.com/mcp, then run /mcp, select Jellypod and sign in. |
| Codex CLI | Run codex mcp add jellypod --url https://mcp.jellypod.com/mcp. A browser window opens to sign in. |
| Zapier | Add the MCP Client by Zapier with the URL, transport Streamable HTTP, OAuth off, and your API key as the bearer token. |
| n8n | Add an MCP Client node with the URL, transport HTTP Streamable and Bearer Auth, using your API key as the credential. |
| Grok Bot | In Marketplace, add a custom remote MCP plugin with the URL, then sign in with OAuth or use an API key as the bearer token. See the Grok Bot guide for a full walkthrough. |
| Other clients | Add the URL as a remote Streamable HTTP server. Use OAuth if the client supports it, otherwise a bearer API key. |
Developers area in Studio
Open Developers in the Studio sidebar for the Overview, API Keys and Integrations tabs. Integrations has copy-and-paste setup steps for each client. API Keys is visible only to administrators and members with permission to manage API keys.
What it can do
The assistant can:
- List and read podcasts, episodes, videos, shorts and existing voiceovers
- Find saved characters, built-in voices and built-in video styles
- Create and update podcasts and characters
- Generate episodes, videos and shorts from a prompt or a finished script
- Update descriptive metadata on episodes, videos, shorts and voiceovers
- Publish, schedule, reschedule and unpublish episodes
- Send feedback about the connector to Jellypod
See the tool reference for names and arguments, and the agent guide for idempotency, polling and retries.
For episodes, videos and shorts, prompt writes new narration using any attached Sources as reference material. To preserve finished narration word for word, use script and assign its speakers to saved characters.
A client that supports MCP Apps plays content inline. Other clients get a Listen or Watch link.
What it cannot do
- Accept files. Research material is limited to public webpage or YouTube URLs and plain text the client already holds. It does not take file uploads, local paths, private URLs or saved source IDs. If the client can read a file, it can pass the extracted text. To give Jellypod the original file, use Studio or the Jellypod API.
- Delete content. Use Studio.
- Create voiceovers, edit production, or manage the account. Voiceover creation, production editing, analytics, voice cloning, billing, team management and account settings are in Studio or, where documented, the API.
Defaults to know
- New podcasts are private, and publishing an episode does not make its podcast public.
- Episodes stay drafts unless
auto_publishis true, which the assistant should use only when you ask it to publish. - A failed generation cannot be retried in place through MCP. For videos, shorts and voiceovers, use Studio or the API to retry the existing item first. A render-only retry costs no credits; other retries charge only for what they still generate. If you choose to start over through MCP, creating new content is a new credit-bearing action. See what a failed generation costs.
Was this page helpful?