Jyoho MCP

Inspiration

As a marketer, I constantly struggled to understand user sentiment and build compelling messaging based on authentic customer feedback. Traditional market research takes weeks and misses the real conversations happening in online communities where users share honest opinions about products and competitors.

What it does

Jyoho MCP transforms how product marketers and PMs access market intelligence. Through a simple chat interface, users can ask natural language questions like "What are developers saying about AWS Lambda?" and instantly receive AI-powered sentiment analysis from Reddit. It turns scattered online conversations into actionable insights for product decisions and messaging strategy.

How we built it

We built a real-time data pipeline that ingests Reddit conversations through Kafka, processes them via a custom MCP server, and delivers insights through Claude Desktop's chat interface.

Challenges we ran into

Three main technical challenges: 1) Building a Kafka pipeline for data streaming from Reddit, 2) Formatting raw Reddit forum data into structured MCP response formats that Claude can effectively process, and 3) Enhancing the MCP response capabilities to handle a wider range of user questions beyond basic sentiment queries.

Accomplishments that we're proud of

Building a working prototype from scratch that successfully demonstrates the core concept. We created a functional chat UI through Claude Desktop's MCP client, implemented a robust MCP server, and established a Kafka data pipeline that captures authentic customer voice from Reddit.

What we learned

How to build MCP servers and work with the ecosystem. We gained insights of real-time data processing, API integration challenges, and the complexities of sentiment analysis at scale.

What's next for Jyoho MCP

1) Expand data sources by integrating additional forums and platforms beyond Reddit for comprehensive coverage. 2) Enhance insights by improving MCP responses with richer metadata, trend analysis, and deeper sentiment breakdowns. 3) Scale broader product coverage beyond demo products to support sentiment analysis for any product or technology. 4) Add advanced features including real-time alerts, competitive benchmarking, and sentiment visualization tools

Built With

• api
• claude
• express.js
• javascript
• kafka
• mcp
• modelcontextprotocol
• reddit
• typescript