Connect Podbrain to Claude Code
Claude Code talks to remote MCP servers over HTTP, so adding Podbrain takes one command in your terminal. Once it is registered, Claude can search about 170 shows and 2,400 episodes, read professional notes, and pull the books a guest recommended, all from inside a coding session.
What you need
- Claude Code installed and signed in. Anthropic's docs say Claude Code needs a Pro, Max, Team, Enterprise, or Console account; the free claude.ai plan does not include it. A third party provider such as Amazon Bedrock also works.
- A terminal. Any folder works, but the default scope ties the server to the folder you run the command in, so the steps below use user scope to make Podbrain available in every project.
- A Podbrain account, created when you sign in from Claude Code, and the free trial. The trial needs a card but charges nothing for your first 30 episodes.
Set up Podbrain in Claude Code
https://podbrain.app/mcp-
Check that Claude Code is installed
Open a terminal and print the version. Run this in your shell, not inside a
claudesession.claude --version -
Register the Podbrain server
Add the server with the HTTP transport. The name
podbrainis just a label; Claude Code uses it to tag tool calls and in later commands. There is no client ID or secret to paste, because Podbrain lets Claude Code register itself when you sign in.claude mcp add --transport http --scope user podbrain https://podbrain.app/mcp -
Confirm the connection
List your servers. The Podbrain row may read
✔ Connectedrather than! Needs authentication, because the server lists its tools without a token. Tool calls still need you to sign in, which is the next step.claude mcp list -
Sign in to Podbrain
Run the login command. Your browser opens the Podbrain sign in page; approve it and return to the terminal. If you work over SSH, add
--no-browserand paste the redirect URL back when asked.claude mcp login podbrainWe have not confirmed that every Claude Code version offers this sign in for a server that already reports Connected. If the command says there is nothing to authenticate, check the server menu in
/mcpfor an Authenticate option. Your first tool call will also ask you to sign in. -
Look at the tools in /mcp
Start a session with
claude, type/mcp, and selectpodbrain. You should see search_episodes, get_episode_notes, get_references, follow_show, and list_follows.claude /mcp -
Ask your first question
Ask about a show in plain words. If Claude Code asks for permission the first time Claude calls a Podbrain tool, approve it. The answer comes back with a link to the episode page.
Try it
What did Huberman Lab say about sleep and caffeine? Use Podbrain and cite the episode.
Claude searches the catalogue for matching Huberman Lab episodes, reads the notes for the best match, and answers with the relevant points plus a link to the episode on podbrain.app. You can swap in The Tim Ferriss Show or The Diary Of A CEO, or ask for the books a guest mentioned.
Turn it off
To remove Podbrain completely, run the command below in your shell. Claude Code also deletes the sign in tokens and client registration it saved for the server. If you added it at another scope, add --scope local or --scope project.
claude mcp remove podbrain --scope user To pause it without deleting anything, open /mcp in a session and turn the server off; Claude Code keeps the entry and marks it disabled. To sign out but keep the server, run claude mcp logout podbrain or choose Clear authentication in the server's /mcp menu.
Questions
Do I have to sign in before Claude can use Podbrain?
Yes. Every tool needs a signed in Podbrain account. If you are signed out, Claude gets a message asking you to sign in. If you have no plan yet, it gets your personal link to start the free trial, which covers your first 30 episodes. After that it is $9.99 a month on Unlimited.
Which scope should I use?
User scope makes Podbrain available in all your projects and keeps it private to you. Local scope, the default, limits it to the current project. Project scope writes a .mcp.json file you can commit, and teammates are asked to approve the server the first time they open the project.
I already added Podbrain as a connector on claude.ai. Do I need this?
Probably not. Anthropic's docs say connectors you add on claude.ai load in Claude Code automatically when you are signed in to Claude Code with that claude.ai subscription account. They do not load when Claude Code runs on an API key. If you also add it with claude mcp add, Claude Code treats the two as the same endpoint and connects once, preferring your local definition.
Claude did not use Podbrain for my question. What now?
Claude picks tools on its own, and a general question can go to web search instead. Mention Podbrain in the prompt, as in the example above. If the tools still do not appear, run claude mcp get podbrain to see the status and any error detail.
The browser did not open when I signed in.
Copy the URL Claude Code prints and open it yourself. If the redirect fails after you approve, paste the full callback URL from the address bar back into the terminal prompt.
Podbrain