---
title: Medium Agent
emoji: 📚
colorFrom: indigo
colorTo: gray
sdk: docker
pinned: true
app_port: 7860
license : mit
short_description: The Intelligence Layer for Medium.com (MCP + Gradio).
tags:
- building-mcp-track-consumer
- building-mcp-track-creative
- agents
- mcp
- medium
- scraping
- audio
- productivity
- teammate is "hetanshwaghela"
---
[](https://huggingface.co/spaces/T0X1N/Medium-MCP)
[](https://opensource.org/licenses/MIT)
[](https://www.python.org/downloads/)
[](https://www.docker.com/)
---
## 📖 The Story: Why Medium Agent?
We live in an era of **Information Overload**. Medium.com is one of the world's richest repositories of human knowledge, but accessing that knowledge is often linear, fragmented, and time-consuming.
**Medium Agent** was built to solve this. It is not just a scraper; it is an **Autonomous Research System**.
By combining **Agentic Workflows**, **Neural Audio**, and the **Model Context Protocol (MCP)**, we have transformed Medium from a static reading platform into a dynamic, queryable, and listenable knowledge base.
> *"Don't just read the internet. Converse with it."*
---
## ✨ What's New (v3.0 - Latest Updates)
### 🚀 MCP Server v3.0
- **12 Tools** (2 new): `medium_related`, `medium_recommend`
- **4 Resources** (1 new): `medium://stats` - server capabilities endpoint
- **9 Prompts** (3 new): `analyze_trending`, `deep_research`, `content_repurpose`
- **Full MCP Annotations**: All tools have `readOnlyHint`, `openWorldHint`, `idempotentHint`
- **Progress Notifications**: Real-time progress for batch operations
- **MCP Logging**: Proper `ctx.info()` and `ctx.warning()` instead of print()
- **HTTP Transport**: Run as HTTP server with `--http` flag
- **Server Instructions**: Comprehensive guidance for LLM clients
### 🎨 UI Enhancements
- **🏠 Hero Tab**: Beautiful landing page with feature overview and gradient design
- **🎧 Improved Audio**: Groq Llama for summarization with natural sentence endings
- **📊 Enhanced Intelligence**: Streamlined analyst reports with PDF export
---
## 🏆 Project Highlights
| Category | Implementation |
| :--- | :--- |
| **🎨 Design / UX** | **"Project Aether"**: Glassmorphic interface with Hero landing page, centered navigation, and premium dark theme. |
| **🤖 Functionality** | **Hybrid Architecture**: Seamless integration of **Gradio** for web and **MCP** for Claude Desktop. |
| **💡 Creativity** | **Sonic Mode**: Transform articles into podcasts with neural TTS (ElevenLabs + OpenAI fallback). |
| **🌍 Impact** | **Democratizing Knowledge**: Research tools for analysts, developers, and visually impaired users. |
---
## 🎥 The Experience (Demos)
* **[Google Drive Folder](https://drive.google.com/drive/folders/1TLX_qLNBd6jU0S_BfZB7sfy4PPGK4Vvf?usp=drive_link)**: Access all project assets and demos.
### 1. 🤖 Claude Desktop Integration
*Watch how Medium Agent gives Claude real-time web browsing capabilities.*
### 2. 🔍 Discover & Search
*Find the signal in the noise with visual masonry-style results.*
### 3. 🎧 Sonic Mode
*Turn your reading list into a playlist. Perfect for commuters.*
### 4. 🧠 Intelligence Dashboard
*Generate professional "State of the Union" reports in seconds.*
---
## 🧠 How It Works
The system operates as an autonomous research pipeline:
1. **User** asks a question via the **Project Aether UI**
2. **Scout Agent** searches DuckDuckGo & RSS for fresh signals
3. **Reader Agent** extracts clean text, bypassing paywalls
4. **Analyst Agent** (Groq Llama 3.3) synthesizes professional reports
5. **Sonic Agent** (ElevenLabs) converts content to podcast audio
---
## 🛠️ Tech Stack
| Technology | Purpose |
| :--- | :--- |
| **[ElevenLabs](https://elevenlabs.io/)** | Neural TTS for Sonic Mode |
| **[OpenAI](https://openai.com/)** | TTS fallback & embeddings |
| **[Groq](https://console.groq.com/)** | Primary LLM (fastest inference) |
| **[Google Gemini](https://deepmind.google/technologies/gemini/)** | Backup LLM + Vision/Embeddings |
| **[Gradio](https://gradio.app/)** | Project Aether web interface |
| **[Playwright](https://playwright.dev/)** | Headless browser scraping |
---
## 🌍 Real-World Impact
**For Researchers & Analysts**
> *Problem*: "I need to know the latest trends but don't have time for 50 articles."
> *Solution*: Analyst Dashboard generates a comprehensive report in 30 seconds.
**For Developers & Commuters**
> *Problem*: "I want to learn about Rust 1.80 while driving."
> *Solution*: Paste the URL into Sonic Mode and listen as a podcast.
**For Accessibility**
> *Problem*: "Most web content is visual-first."
> *Solution*: High-quality audio makes Medium accessible to everyone.
---
## 📚 Documentation
* **[👉 GRADIO_APP.md](GRADIO_APP.md)**: Design philosophy and UI features
* **[🔌 MCP_SERVER.md](MCP_SERVER.md)**: Claude Desktop integration guide
---
## 🚀 Quick Start
### Docker (Recommended)
```bash
git clone https://huggingface.co/spaces/T0X1N/Medium-MCP
cd Medium-MCP
docker build -t medium-agent .
docker run -p 7860:7860 --env-file .env medium-agent
```
### Local Python
```bash
pip install -r requirements.txt
playwright install chromium
python app.py
```
---
## 📢 Connect
> **"Bridging the gap between a sleek UI and Agentic AI."**
[](https://www.linkedin.com/posts/nikhil-pravin-pise_bridging-the-gap-between-a-sleek-ui-and-agentic-share-7400935316854943744-bGfs)
---
## 🙏 Acknowledgments
Built with ❤️ for the **Open Source AI Community**.
* **Groq**: Lightning-fast LLM inference
* **Google DeepMind**: Gemini for Vision/Embeddings
* **Anthropic**: MCP Standard
* **Hugging Face**: Platform & Infrastructure
* **ElevenLabs**: Neural Voice Technology