In one of my recent conversations with an executive leadership (Sree G) in engineering, who expressed his appreciations for my blog posts and wished to pick some of his favorite posts on a few topics of his interest - Gen AI, AWS Cloud etc., to get a summary of my thoughts about it.
Post the conversation, I wanted to give it a shot myself and first looked for existing solutions but couldn't find one. I felt it was an idea worth trying out. A few weeks goes by, and I publish my MVP version of Query Blogger MCP Server project to tackle this problem. Do check it out, download it, connect it with your LLM, and fire your questions against my Blog as knowledge base to get your queries answered.
Conversation >> Idea >> PoC >> Commodity (aka Product)
The milestone is achieved with the MVP and the journey continues. For there are a lot of things to cover the ground that is expanding. One of the things that I relish in my life is that when I get done with something, it leaves me with a lot of questions and finding answers to those lead me to another set of questions and it goes on keeping me ever occupied.
Here are some of the questions that I as an avid Gen AI enthusiast is seeking answers for and expect the questions to only grow as I find answers to a few of them:
- How do I quickly test to validate my product as a Product Manager to quench my thirst - integrate with existing LLMs like Claude or OpenAI, or integrate with local LLM?
- How do I quickly test my MCP server as a Tester to validate my understanding of its behavior - are there any UI tools out there to connect to MCP servers like PostMan etc. that we used to connect to REST API servers?
- What does it mean to integrate it with various LLMs of different sizes?
- Are there better way to build this MCP server in terms of best-practices to adopt?
- Is there any MCP Spec (like REST) or community guidelines on the contract for response schema?
- What does it mean to have a demo client to show-case this differently?