PodMule

com.podmulePublished by podmule.comOfficial publisherv0.1.0Updated Oct 24, 2026

Search, draft and manage your podcast workspace from your agent.

VerifiedStreamable HTTPWeb executableProductivity & WorkflowKnowledge & MemoryMedia & Design

Overview

AI-generated overview

PodMule connects an assistant to your own podcast workspace to search shows, episodes and transcripts, draft clips and show notes, and run read-only…

What you can do
  • Ask PodMule for concise read-only answers, summaries, diagnostics or advisory route plans
  • Search the capabilities available to your OAuth token to get exact names and schemas
  • Execute one visible capability; reads run directly, while writes stay read-only except an allowlisted update pilot for modern clients
  • Work with shows, episodes, transcripts, the back-catalogue knowledge base, clips, guest CRM and interview bookings through those capabilities
When to use it
Useful when you already run a podcast on PodMule and want an agent to answer questions about your back catalogue, find quotes and timestamps, draft clips or show notes, or brief you on an upcoming guest.
Requirements
A PodMule workspace account. The server is a remote Streamable HTTP endpoint that requires OAuth sign-in through a browser consent screen; no API key is copied. Non-interactive sessions cannot complete the OAuth flow, so authorise once interactively first.
Before you install
The connection reads your workspace data, including transcripts, guest CRM and bookings, and you choose which workspace it can access at consent. Some skills change your workspace and are meant to ask first; the execute tool can run writes, though most remain read-only and dangerous writes are blocked. Treat the consent screen as granting access to real podcast data.

Installation

In SourceWeft

  1. Open PodMule in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/com.podmule%2Fpodmule/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "podmule": {
      "type": "http",
      "url": "https://app.podmule.com/mcp"
    }
  }
}

README

[Image]

PodMule plugin

Connect your agent to your own PodMule podcast workspace — shows, episodes, transcripts, the knowledge base built from your back catalogue, clips, guest CRM and interview bookings.

The plugin bundles two things: a reference to PodMule's remote MCP server, and five skills that tell your agent when to use which part of it. Connecting without the skills gives you three tools and no idea how to drive them.

You need a PodMule workspace. This plugin talks to your data; it does nothing on its own. You sign in with your normal PodMule account the first time your agent connects — there is no API key to copy. If you don't have an account yet, start at podmule.com.

Connector URL: https://app.podmule.com/mcp · Full setup guide: app.podmule.com/mcp-setup

Install

Everywhere except Claude Code, the whole install is the repo name — Squared-Lemons-Ltd/podmule-plugin — pasted into a Plugins settings screen.

Claude Code

/plugin marketplace add Squared-Lemons-Ltd/podmule-plugin/plugin install podmule@podmule

Then /mcp and sign in. A browser window opens for PodMule's consent screen; if you belong to more than one workspace you pick which one the connection uses.

Non-interactive sessions (claude -p, the Agent SDK) cannot complete the OAuth flow and will report the tools as unavailable. Authorise once from an interactive session first.

Claude Desktop, ChatGPT and Codex

All three take the repo name directly. Open the Plugins section, choose Add custom, and enter:

Squared-Lemons-Ltd/podmule-plugin

Then install PodMule and sign in. One package, read by all of them.

The ChatGPT chat surface is a separate thing and is not reachable this way — it needs a connector ID minted by your own developer-mode registration, which no distributable package can carry.

One click, other surfaces

[Add to Cursor] [Add to VS Code] [Add to VS Code Insiders] [Add to LM Studio] [Add to Replit] [Add to Goose]

These add the MCP connection only — the skills come with the plugin package, so Claude Code and ChatGPT desktop get more out of it.

Anything else

Any MCP client that speaks Streamable HTTP with OAuth can point at the URL directly:

json
{  "mcpServers": {    "podmule": {      "type": "http",      "url": "https://app.podmule.com/mcp"    }  }}

"type" is not optional. Claude Code reads a url entry without it as a stdio server, skips it, and errors.

What the skills do

SkillAsk it
find-in-back-catalogue"Which episodes discuss burnout?" · "Find the exact quote and timestamp."
draft-clips"Draft clips from Tuesday's episode."
interview-brief"Brief me on my next guest."
episode-writeup"Generate show notes and titles for episode 42."
workspace-triage"What needs my attention today?"

find-in-back-catalogue is read-only and the best first thing to try. The clip and write-up skills change your workspace, so they ask before doing so.

Repo layout

.claude-plugin/marketplace.json    the catalog, read by both Claude and ChatGPT desktopplugins/podmule/  .claude-plugin/plugin.json       Claude manifest  .codex-plugin/plugin.json        Codex/ChatGPT manifest  .mcp.json                        the connector declaration, shared by both  skills/                          five skills, shared by both  assets/                          brand mark, taken from the PodMule marketing siteserver.json                        the official MCP registry entryscripts/check-manifests.mjs        structural checks on the manifestsscripts/check-capability-drift.mjs fails when a skill names a capability that no longer exists

Both manifests point at one skills directory and one .mcp.json, so the two ecosystems can never drift apart in content — only in the presentation fields each one reads.

Checks

node scripts/check-manifests.mjsnode scripts/check-capability-drift.mjs

The drift check reads PodMule's public capability list and fails if a skill mentions something the server no longer registers. It needs no credentials, and runs in CI on every push and pull request. Renaming a capability in the product is exactly the kind of change that otherwise breaks installed guidance silently.

Licence

MIT — see LICENSE.

Source: README.md at commit 84c694a

Tools

3

ask_podmule

Ask PodMule for a concise read-only domain answer or an advisory route plan. Use this for quick questions, summaries, diagnostics, drafting, "what should I do?" prompts, and ambiguous action requests. This tool never performs writes. For exact capability contracts use search; for one selected action (read or write) use execute.

Parameterscontextrequestrequiredresponse_format

execute

Execute one OAuth-visible PodMule capability. Reads run directly. Modern MCP 2026 clients with MRTR support may preview and affirm the allowlisted update_episode pilot; legacy clients and every other MCP write remain read-only. Dangerous writes are always blocked — do those in the PodMule app. Use ask_podmule for ambiguous requests and search for exact schemas.

Parametersnamerequiredinput

Version history

1
  1. v0.1.0LatestSep 16, 2026

Score

100% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listedRequired15 pts
  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts