For agent builders and AI-first teams
The podcast intelligence MCP server for your AI
Let Claude, ChatGPT, or Cursor ask who is advertising where — and get back timestamped ad reads, promo codes, and standing brand watches. One connection, no scraping.
Just ask, in plain English
“Which brands advertised on the top health podcasts last month?” Your agent picks the right tools and comes back with timestamped ad reads — no query syntax, no dashboard.
7,500 ad-carrying podcasts
Not a scrape of everything. A ranked cohort of the shows where advertising money actually moves, so your agent's answers are about shows worth caring about.
Connect once with a link
One HTTP endpoint with your key in the URL. No cloning a repo, no local install, no config file surgery. Paste it into your client and start asking.
Every ad read timestamped
Sponsors, promo codes, and ad copy come back anchored to the exact second in the transcript, so your agent can cite its evidence instead of asserting it.
Standing brand watches
Your agent doesn't just read — it can create a monitor that keeps running. Ask it to watch a competitor, and you get told when that competitor starts advertising.
Stays in your account
Anything an agent creates lands in your PodZeus workspace for review. You see what it set up, and you can edit or kill it.
What your agent can ask for
No query language. These are complete prompts — the agent works out which tools to call and in what order.
“Which brands advertised on the top 20 health podcasts last month, and what promo codes did they use?”
“Watch for any mention of our biggest competitor and email me the moment they start advertising.”
The tools it gets
Shaped as a loop an agent can walk on its own: find the shows, ask what we hold on them, query, then persist a monitor. The coverage step is the one that matters — an agent that can ask where the data ends doesn't invent the rest.
| Stage | Tool | Does |
|---|---|---|
| Discover | listTrackedShows | Browse the ranked cohort, filtered by category, size, or region. |
| Constrain | getShowCoverage | Ask what we actually hold for a show before querying it: is it in the cohort, what date range, how many episodes, are sponsors extracted. Off-cohort shows return the nearest tracked alternatives rather than an empty result. |
| Act | searchTranscripts | Full-text search across analyzed episodes, returning timestamped context. |
| Act | getSponsors | Every advertiser on a given show or episode, with ad copy and promo codes. |
| Act | getAdvertiser | Flip it around: every show a given brand has advertised on, over time. |
| Write | createBrandWatch | Set up a standing monitor on a brand, competitor, or keyword. |
| Verify | listBrandWatches | What monitors are running, and what they have caught. |
How you'll connect
Your API key goes in the URL. There is nothing to install and no OAuth round trip — paste one line and your agent has the tools.
Claude Code
One line in your terminal:
claude mcp add --transport http podzeus https://api.podzeus.com/mcp/YOUR_API_KEYClaude desktop / Cursor
Add a remote MCP server in settings pointing at:
https://api.podzeus.com/mcp/YOUR_API_KEYChatGPT
Settings → Connectors → Add custom connector, then paste:
https://api.podzeus.com/mcp/YOUR_API_KEYAnything else
Any MCP client that speaks streaming HTTP, with a bearer token instead:
Authorization: Bearer YOUR_API_KEY → https://api.podzeus.com/mcpFrequently asked questions
Start discovering podcast insights today
Start with a 7-day trial and explore a growing catalog of popular podcasts. No credit card required.
No credit card required • 7-day trial • Cancel anytime