podcast-transcript-fetcher

Category: Research · Version: 1.1.0 · Author: farizanjum · Compatibility: All targets (claude, codex, opencode, gemini, anti-gravity, openclaw, hermes)

Fetch, search, and batch-transcribe transcripts from 5 top podcasts. Tier 2 (RSS+Groq Whisper) is the recommended approach — fast, free, and the most reliable. Tier 1 free sources are best-effort; Tier 3 Taddy API is the commercial/premium option.

Supported podcasts: Lenny’s Podcast, Dwarkesh Podcast, Cheeky Pint, 20VC (20 Minutes VC), A16z Podcast

Install

npx CLI (Recommended)
skills.sh
Claude Code Native
Claude Desktop
Manus AI

npx "@opendirectory.dev/skills" install podcast-transcript-fetcher --target claude

npx skills add Varnan-Tech/opendirectory --skill podcast-transcript-fetcher

/plugin marketplace add Varnan-Tech/opendirectory
/plugin install opendirectory-gtm-skills@opendirectory-marketplace

Copy https://github.com/Varnan-Tech/opendirectory/tree/main/skills/podcast-transcript-fetcher into download-directory.github.io to download the ZIP.
Open Claude Desktop -> Customize -> Skills -> + -> Upload a skill and drop the folder containing SKILL.md.

Quick Start

# Set up Python dependencies
pip install requests python-dotenv

# Copy and fill in your API key
cp .env.example .env
# Then edit .env with your GROQ_API_KEY

# Get the latest Lenny's Podcast transcript
python scripts/get_transcript.py "Lenny's Podcast" --latest

# Search across ALL podcasts by keyword
python scripts/get_transcript.py --search "Marc Andreessen"

# Batch-transcribe last 3 episodes
python scripts/get_transcript.py "Dwarkesh Podcast" --last 3

Transcription Methods

Method	Cost	Speed	Quality	Setup	Recommendation
Free sources	$0	Instant	Varies	None	Best-effort, limited availability
Groq Whisper (Tier 2)	Free tier	~10s/hr audio	2.7% WER	`pip install groq` + API key	RECOMMENDED
Local faster-whisper	$0 (compute)	~1.5x realtime	2.7% WER	`pip install faster-whisper`	Offline alternative
Taddy API (Tier 3)	$75/mo+	Instant	High	`TADDY_API_KEY`	Premium / commercial

Prerequisites

Python 3.10+
ffmpeg — for audio compression (Groq’s 25 MB limit)
- Windows: winget install ffmpeg or scoop install ffmpeg
- macOS: brew install ffmpeg
- Linux: apt install ffmpeg
Groq API key — free at console.groq.com
Works with: Claude Code · OpenCode · Codex CLI · Gemini CLI · Cursor

Usage

Single Episode

# Get latest episode (auto-detects best method)
python scripts/get_transcript.py "Lenny's Podcast" --latest

# Specific episode by title
python scripts/get_transcript.py 20vc --episode "Sam Altman"

# Force a specific method
python scripts/get_transcript.py a16z --latest --method whisper

# Save to file
python scripts/get_transcript.py lennys --latest --output episode.md

Cross-Podcast Search

# Find episodes by keyword across all 5 podcasts
python scripts/get_transcript.py --search "regulation"

# Search by guest name
python scripts/get_transcript.py --guest "Marc Andreessen"

# Filter to one podcast
python scripts/get_transcript.py "Lenny's Podcast" --search "vibe coding"

Batch Transcription

# Transcribe last 5 episodes
python scripts/get_transcript.py "20vc" --last 5

# Search + auto-transcribe top matches
python scripts/get_transcript.py --search "AI safety" --transcribe --transcribe-count 5

How It Works

User provides podcast + episode
      ↓
Tier 1: Free sources (best-effort, limited availability)
  → Lenny's: GitHub archive (269 transcripts)
  → Others: website scrape, Substack PDF
  → Found? Return instantly
  → Not found? Fall through to Tier 2
      ↓
Tier 2: RSS + Whisper transcription [RECOMMENDED]
  → Download MP3 from podcast RSS feed
  → Compress if >25 MB (ffmpeg)
  → Transcribe via Groq Whisper API (free tier, ~10s/hr audio)
  → Fast, free, and works for every podcast
      ↓
Tier 3: Taddy API (commercial/premium)
  → Requires TADDY_API_KEY ($75/mo+)
  → Use for large-scale or production needs

Search and batch features build on Tier 2:

Search scans RSS metadata (titles + descriptions) across all podcasts — instant, no API key needed
Batch downloads and transcribes the last N episodes in sequence
Pipeline chains search results into batch transcription

Project Structure

podcast-transcript-fetcher/
├── SKILL.md                ← AI instructions (the brain)
├── README.md               ← This file
├── package.json            ← Skill metadata
├── .env.example            ← Template for API keys
├── scripts/
│   ├── get_transcript.py   ← Core CLI (3-tier transcript fetcher)
│   └── podcasts.json       ← Podcast registry (RSS feeds, sources)
└── references/
    └── podcasts.md         ← Detailed source documentation

Use with AI Agents

Once you have a transcript, ask your AI agent to analyze it:

“Summarize this transcript from Lenny’s Podcast” “Extract 3 actionable insights from this 20VC episode” “Compare the Dwarkesh and A16z takes on AI safety” “Find all episodes where they discuss [topic]“

Contributing

Want to add more podcasts? Open a PR or issue at the OpenDirectory repo. To add a podcast, just add an entry to scripts/podcasts.json with the podcast name, RSS feed URL, available transcript sources, and a slug. See the existing entries for the format.

Catalog

Visual & Media

Content

Launch

GTM Intelligence

Outreach

Research

Developer Tools

Other

podcast-transcript-fetcher

Install

Quick Start

Transcription Methods

Prerequisites

Usage

Single Episode

Cross-Podcast Search

Batch Transcription

How It Works

Project Structure

Use with AI Agents

Contributing

​Install

​Quick Start

​Transcription Methods

​Prerequisites

​Usage

​Single Episode

​Cross-Podcast Search

​Batch Transcription

​How It Works

​Project Structure

​Use with AI Agents

​Contributing

Install

Quick Start

Transcription Methods

Prerequisites

Usage

Single Episode

Cross-Podcast Search

Batch Transcription

How It Works

Project Structure

Use with AI Agents

Contributing