HN Thread Link: https://news.ycombinator.com/item?id=47037501

   Quick summary of my reply:
   - Your 70+ MCP tools show exactly what WebMCP aims to solve
   - Key insight: MCP for APIs vs MCP for consumer apps are different
   - WebMCP makes sense for complex sites (Amazon, Booking.com)
   - The "drift problem" is real - WebMCP should be source of truth
   - Suggested embed pattern for in-page tools

I just open-sourced *Agentic CI* - an intelligent CI system that uses LLMs to actually understand your code changes.

*What it does:*

*Semantic Analysis* - Understands what changed at a code level, not just file level. Identifies risk areas and affected components.

*Intelligent Test Selection* - Predicts which tests to run based on changes. No more running 2,000 tests for a 1-line change.

*Failure Root Cause Analysis* - Explains why tests failed with actual root causes, not just stack traces. Suggests fixes.

*Pipeline Optimization* - Tracks flaky tests, auto-quarantines problematic ones, and suggests parallelization strategies.

*Tech:* Python, FastAPI, supports Ollama (local), OpenAI, and Anthropic.

Inspired by Peter Steinberger's vision for intelligent CI systems.

100% open source under MIT license.

GitHub: https://github.com/yksanjo/agentic-ci

#DevTools #CICD #LLM #OpenSource #SoftwareEngineering

  A free, open-source investor database with 6,500+ investors including:

  - 217 Family Offices
  - 500+ Investment Companies
  - 78 Venture Capital Firms
  - 1,000+ AI-focused investors

  Built using:
  - SEC EDGAR public filings
  - Next.js 16 & React 19
  - Supabase (PostgreSQL)
  - Deployed on Vercel

  Perfect for founders researching potential investors for their next raise.

  Live demo: https://vc-intelligence-mcp.vercel.app
  GitHub: https://github.com/yksanjo/vc-intelligence-mcp

  #startup #venturecapital #opensource #nextjs #fundraising

  Hi HN,

  I built MCP Discovery API, an agent-native discovery layer for Model Context Protocol servers.

  The problem: AI agents can only use tools you pre-configure. They can't discover new capabilities dynamically.

  The solution: A semantic search API that returns the best MCP server for any task.

  Example:
  Query: "I need to send emails with templates"
  Response: Gmail MCP server (87% confidence) + install: npx -y @anthropic/mcp-server-gmail

  Features:
  - Semantic search using OpenAI embeddings + pgvector
  - Real-time performance metrics (latency, uptime)
  - Side-by-side server comparison
  - 24 servers indexed, 91 capabilities tagged

  Tech: TypeScript, Supabase, Railway, MCP SDK

  Live: https://mcp-discovery-production.up.railway.app
  GitHub: https://github.com/yksanjo/mcp-discovery

  Feedback welcome!

I built a small web app that lets you upload or record a cat’s meow and get a simple analysis of what the cat might be expressing (attention, stress, hunger, etc.). This started as a fun experiment with audio feature extraction + lightweight ML, but it turned into a surprisingly entertaining way to think about non-human vocal patterns. The app runs entirely in the browser (no account required). I’d love feedback on: whether the analysis feels believable or useful ideas for better features or datasets ethical / scientific limitations I should be clearer about