spokenmd/spoken

Overview

Spoken is a JavaScript API that converts published podcast episodes into clean Markdown transcripts with accurate speaker attribution. It is designed for AI agents to programmatically access and process podcast content.

Best for

Best for
Developers building AI agents that need to ingest and process podcast transcripts programmatically.

Use cases

• Extract structured transcripts from podcast episodes for analysis
• Feed podcast content into AI agents for summarization or search
• Build applications that index or search across multiple podcast shows

How to use

Tools exposed

• search_podcasts
• list_episodes
• get_transcript
• get_balance

Tested with

Claude Desktop, Cursor, Cline

Example client config

{\n  "mcpServers": {\n    "spoken": {\n      "command": "npx",\n      "args": ["-y", "spoken-mcp"],\n      "env": { "SPOKEN_API_KEY": "pt_your_key" }\n    }\n  }\n}

Notes

Spoken is a JavaScript API that converts published podcast episodes into clean Markdown transcripts with accurate speaker attribution. It is designed for AI agents to programmatically access and process podcast content.

2 stars on GitHub. Last updated 2026-06-05. Licensed MIT.

Use cases

• Extract structured transcripts from podcast episodes for analysis
• Feed podcast content into AI agents for summarization or search
• Build applications that index or search across multiple podcast shows

Pros

• Outputs clean Markdown with real speaker names, not generic labels
• Simple API focused on a single, well-defined task
• Open source with a permissive license

Cons

• Limited to published podcasts only, no live or private audio
• Small community (2 stars) suggests early stage or niche adoption
• No built-in support for non-English or heavily accented speech

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.