📺 Watch the full tutorial on YouTube

🎬 Zero-API YouTube-to-Blog SEO Engine — Agent Reach + Ollama

Turn any YouTube video into a fully structured, SEO-optimized blog post - locally, for free, in seconds.

Features · Tech Stack · Setup · Files · Use Cases · Future Ideas

---

🌟 What is This?

A zero-cost, privacy-first Streamlit app that:

1. Takes any YouTube URL as input
2. Uses Agent Reach (yt-dlp) to download the video transcript - no YouTube API key required
3. Strips timestamps with regex to produce clean plain text
4. Sends the transcript to a local Ollama LLM via its REST API
5. Outputs a fully structured, SEO-optimized blog post with meta description, headers, and bullet points

💡 No API costs. No rate limits. No data leaves your machine.

---

✨ Features

Feature	Description
🔑 Zero-API	No YouTube API key - uses yt-dlp (148K ⭐) via Agent Reach methodology
🦙 Local LLM	Powered by Ollama - runs 100% offline after setup
📝 SEO Output	Auto-generates meta description, H2/H3 headers, bullet points
🌊 Streaming	Live streaming response directly in the Streamlit UI
🧹 Auto Cleanup	Temporary .vtt transcript files deleted after use
⚡ Fast	Uses qwen2.5-coder:1.5b for quick, sharp responses

---

🛠️ Tech Stack

Layer	Tool	Why
UI	Streamlit	Instant Python web UI, zero frontend code
Transcript	yt-dlp via subprocess	Agent Reach's primary YouTube backend - no API key, 1800+ sites
LLM	Ollama	Run powerful models locally, fully free
Model	qwen2.5-coder:1.5b	Fast, efficient, great at structured writing
HTTP	requests	POST to Ollama's local REST API
Parsing	re (regex)	Strip VTT timestamps cleanly
Cleanup	os	Remove temp .vtt files post-extraction

---

⚙️ Setup Instructions (Windows PowerShell)

1. Clone This Repo

git clone https://github.com/47thtechcorner/RayCodes_AgentReachYT.git
cd RayCodes_AgentReachYT

2. Install Agent Reach (enables yt-dlp YouTube access - the core of this project)

⚠️ Agent Reach is the main tool that powers transcript extraction. You must clone and install it separately from its own repository.

# Install Agent Reach from source
pip install https://github.com/Panniantong/agent-reach/archive/main.zip

# This auto-installs yt-dlp and all YouTube channel dependencies
agent-reach install --env=auto

What this does: Agent Reach designates yt-dlp as its YouTube backend. Running agent-reach install installs and health-checks it for you - no YouTube API key, no rate limits.

3. Install Python Dependencies

pip install streamlit requests

4. Pull the Ollama Model

# Make sure Ollama is running (https://ollama.com)
ollama pull qwen2.5-coder:1.5b

5. Run the App

streamlit run app.py

Open http://localhost:8501 in your browser and paste any YouTube URL.

---

📁 Project Files

RayCodes_AgentReachYT/
├── app.py              # Main Streamlit application (≤150 lines)
└── README.md           # This file

app.py

The entire application in a single, minimal Python file:

Section	Lines	Purpose
get_transcript(url)	~20	Calls yt-dlp via subprocess to download .vtt subtitles
parse_vtt(filepath)	~15	Reads file, strips timestamps with regex, returns clean text
generate_blog(text)	~30	Sends text + system prompt to Ollama API, streams response
main() / Streamlit UI	~30	URL input, button, spinner, live markdown output

---

💡 Use Cases

• Content Creators - Repurpose your YouTube videos as blog posts automatically, boosting SEO without extra writing effort
• Marketers - Extract insights from competitor YouTube videos and publish keyword-rich summaries
• Educators - Convert lecture videos or tutorials into readable, shareable study notes
• Developers - Prototype AI-powered content pipelines locally with zero cloud costs
• Journalists / Researchers - Rapidly digest long-form interview or podcast videos into structured article drafts

---

🔮 Future Ideas

• Multi-language support - Auto-detect and translate non-English VTT transcripts before blogging
• Batch mode - Process entire YouTube playlists (channel-level SEO campaigns)
• CMS export - One-click publish to WordPress, Ghost, or Hashnode via their REST APIs
• Custom prompts - Let users pick blog style: listicle, deep-dive, news article, or tutorial
• Keyword injector - Input target SEO keywords and have the LLM weave them naturally into the post

---

🧠 Agent Reach Methodology

This project is built on the Agent Reach philosophy:

"The most reliable access path for each platform - chosen, installed, and health-checked for you."

For YouTube, Agent Reach designates yt-dlp as the primary backend - zero API key, zero rate limits, 1800+ supported sites. We invoke it via subprocess directly, exactly as an AI Agent would.

YouTube URL → yt-dlp (subprocess) → .vtt file → regex parse → clean text → Ollama LLM → SEO Blog ✅

---

📄 License

MIT © 2026 - Free to use, fork, and build upon.

---

_{agent reach tutorial · yt-dlp no api key · youtube to blog ai · ollama local llm · streamlit ai app · free content engine · ai blog generator · youtube transcript extractor · local ai content automation · qwen2.5 ollama · zero api youtube scraper · ai seo blog writer · python youtube automation · agent reach yt-dlp setup · free ai content pipeline}