PodPast.ai for Claude Pro Users: Your Podcast Library as a Native Claude Tool

Workflow 1: MCP setup and first query in under five minutes

The MCP connection is designed to take as little friction as possible. Sign up at PodPast.ai, add your first podcast feed (paste an RSS URL or YouTube channel link), and wait a few minutes for the first episodes to transcribe. Then open your account dashboard and copy your personal MCP URL — it looks like https://api.podpast.ai/mcp/YOUR_TOKEN.

Open Claude Desktop, go to Settings → Developer → Edit Config, and add the PodPast server to your claude_desktop_config.json:

{
  "mcpServers": {
    "podpast": {
      "type": "http",
      "url": "https://api.podpast.ai/mcp/YOUR_MCP_TOKEN"
    }
  }
}

Restart Claude Desktop. You will see a hammer icon in the chat toolbar confirming active MCP tools. Type "Search my pod for [any topic from a podcast you added]" and Claude will call the search_pod tool, retrieve ranked transcript chunks from your vault, and present them with timestamps.

From that moment on, every Claude conversation has access to your podcast library. You do not need to paste transcripts, manage context windows, or think about retrieval. Claude handles it natively via the MCP tools whenever podcast knowledge is relevant to your query.

Workflow 2: Asking Claude synthesis questions spanning dozens of episodes

The most powerful capability unlocked by the MCP connection is synthesis across your entire vault. Instead of searching for specific passages, you ask Claude a high-level question and Claude uses the ask_pod tool to retrieve the most relevant content from across all your feeds and synthesise a grounded answer.

Examples of synthesis queries that become possible with PodPast MCP:

• –"What are the most common arguments for and against [policy position] across the podcasts I follow?"
• –"How has [expert's name] view on [topic] evolved across their podcast appearances?"
• –"Summarise everything my pod contains about [technical concept], with sources."
• –"What did the founders on [podcast series] say about hiring in the early stages?"
• –"Find any contradictions in what [two different experts] have said about [topic]."

Claude retrieves the relevant passages using PodPast's vector search, then synthesises an answer that includes the source episode and timestamp for every claim. The output is grounded in your actual podcast library — not general training data — and every claim is verifiable.

This workflow is particularly powerful for knowledge workers who use Claude for research, writing, or decision-making. Instead of treating Claude as a general-purpose AI, you extend it with your specific information domain — the podcasts and creators you find most authoritative — and get answers grounded in that curated corpus.

Workflow 3: Using Claude Projects with your podcast vault

Claude Projects lets you pre-configure a persistent context for a specific research area, project, or role. With PodPast's MCP server active, your podcast vault is available as a tool in any Project conversation — meaning every chat within that Project can draw on your full transcript library.

A practical setup: create a Claude Project for your primary research or work area. Add a system prompt that instructs Claude to use your PodPast vault when answering domain questions. Now every conversation in that Project automatically has access to your podcast library as a background reference — without you needing to prompt for it explicitly.

This is particularly useful for ongoing research projects that span weeks or months. As new podcast episodes publish and are automatically indexed by PodPast, your Project's available knowledge base grows without any configuration changes. The Project stays current with the latest commentary from the creators you follow.

Claude Code users can also benefit: if your work involves technical topics discussed in developer podcasts, engineering leadership shows, or startup operations content, the PodPast MCP makes that content available mid-session in Claude Code. Ask for a code review approach and Claude can draw on relevant podcast discussions of engineering practices alongside its training knowledge.

What PodPast adds to Claude

• ✓search_pod — semantic search returning ranked transcript chunks with timestamps
• ✓ask_pod — Claude asks for a synthesised answer; PodPast retrieves and Claude composes with citations
• ✓add_to_pod — add new RSS feeds or YouTube channels mid-conversation without leaving Claude
• ✓get_pod_info — check your vault statistics and transcription credit balance from Claude
• ✓No context limit — RAG-based retrieval means unlimited vault size without hitting context windows
• ✓Works in Claude Desktop, Claude Code, and Claude Projects
• ✓Vault updates automatically — new episodes available to Claude without any reconfiguration

Questions from Claude power users