Turn one podcast episode into a full week of content — powered by GPT-4o.
One transcript. Five platforms. Zero manual writing. Paste a podcast transcript and get a polished blog post, Twitter thread, LinkedIn post, email newsletter, and YouTube description — each crafted with platform-specific best practices.
| Output | Details |
|---|---|
| 📝 SEO Blog Post | ~1,500 words with H2/H3 structure, guest quotes, meta description, and CTA |
| 🐦 Twitter/X Thread | 10-tweet thread with scroll-stopping hook and engagement tactics |
| 💼 LinkedIn Post | Pattern-interrupt opener, 150–300 words, hashtags, engagement question |
| 📧 Email Newsletter | 3 key lessons with action items, warm conversational tone |
| 🎬 YouTube Description | Timestamps, tags, guest bio, SEO keywords, subscribe CTA |
Each piece uses names, quotes, and specific details from your transcript — not generic filler.
┌──────────────────────────────────────────────────────┐
│ Streamlit UI │
│ │
│ ┌────────────────────────────────────────────────┐ │
│ │ Input: Paste │ Upload .txt │ Sample Transcript│ │
│ └───────────────────────┬────────────────────────┘ │
│ ▼ │
│ ┌─────────────────────┐ │
│ │ Content Pipeline │ │
│ │ │ │
│ │ For each format: │ │
│ │ ┌───────────────┐ │ │
│ │ │ System Prompt │ │ │
│ │ │ + Format- │ │ │
│ │ │ Specific │──┼──► OpenAI GPT-4o │
│ │ │ Template │ │ │
│ │ └───────────────┘ │ │
│ └─────────┬───────────┘ │
│ ▼ │
│ ┌──────────────────────────────┐ │
│ │ Results Dashboard │ │
│ │ │ │
│ │ 📊 Word counts per format │ │
│ │ 📑 Tabbed content viewer │ │
│ │ 📦 Download all as ZIP │ │
│ │ 📋 Individual downloads │ │
│ └──────────────────────────────┘ │
└──────────────────────────────────────────────────────┘
git clone https://github.com/arushs/podcast-repurposer.git
cd podcast-repurposer
pip install -r requirements.txt
streamlit run app.py- Enter your OpenAI API key in the sidebar
- Paste a transcript, upload a
.txtfile, or use the built-in sample - Click ✨ Generate All Content
- Download individual files or grab everything as a ZIP
A sample transcript is included so you can demo it instantly — no podcast required.
Each content type has a dedicated prompt template in prompts.py with platform-specific constraints:
- Blog — enforces H2/H3 structure, 3–5 direct quotes, meta description
- Twitter — 280-char limit per tweet, numbered threading, hook-first
- LinkedIn — short paragraphs, pattern-interrupt opener, hashtags
- Newsletter — lesson + explanation + action item format
- YouTube — timestamps, tags, guest bio, SEO keywords
All prompts share a system prompt that emphasizes using specific details (names, quotes, numbers) from the transcript.
podcast-repurposer/
├── app.py # Main Streamlit application
├── prompts.py # GPT-4o prompt templates per content type
├── sample_transcript.py # Built-in demo transcript
├── requirements.txt # Python dependencies
├── .env.example # Environment variable template
├── .streamlit/
│ └── config.toml # Streamlit theme config
├── LICENSE # MIT License
└── README.md
- Push to GitHub
- Go to share.streamlit.io
- Connect your repo → select
app.py - Add
OPENAI_API_KEYto Secrets - Deploy 🚀
- Streamlit — Interactive UI with progress tracking
- OpenAI GPT-4o — Content generation
- Python 3.9+ — Runtime
MIT — see LICENSE.
Built by Arush Shankar · Powered by GPT-4o