A briefing in your podcast app. Memory you query from Claude.
Podgest listens to the podcasts you subscribe to. You get a spoken briefing in Overcast, Apple Podcasts, or Spotify. Every episode is also embedded into a queryable memory — so from Claude or ChatGPT you can pull relevant content from those podcasts into the context you already keep there. Don’t start over.
01
Pick your feeds, a length between 5 and 20 minutes, and a delivery time. Overnight we transcribe what’s new, write a briefing, and publish it to a private RSS URL. You add that URL once. After that it is just another episode in the app you already open on the walk.
03
Connect mcp.podgest.app to Claude, ChatGPT, or Cursor.
The point is not a new chat window, and it is not their memory
feature. You already keep context somewhere — a fitness project,
a research book, a trading mandate. Podgest remembers the shows.
Those clients are where that memory becomes useful: inside the
thread you already trust, instead of pasting a transcript and
starting from zero.
04
Every episode is embedded into Podgest’s memory. Ask what guests have said about a topic across months. Compare hosts. Pull a timestamp and a listen link. Trends over time — not “summarize this one file.” You reach that memory from Claude or ChatGPT, inside the context you already have there. That’s what makes the connection useful on day 90, not just day one.
05
You paste an Anthropic key for the briefing script and an OpenAI key for speech and embeddings. We handle transcription, the private feed, memory, and MCP. Costs land on your invoices, so you see what a digest actually costs — typically around ~$5 a month. No hidden seat fee.