HN Showcase
Back to browse
GitHub Repository

Pre-build reality check for AI coding agents. Scans GitHub, HN, npm, PyPI, Product Hunt. MCP server. 290+ stars.

708 starsPython

Idea-reality-MCP – MCP server that searches real data before you build

by mnemoxai·Feb 27, 2026·1 point·1 comment
Visit ProjectView on HN

AI Analysis

●●●BangerBig BrainShip ItSolve My Problem

Stops idea-stage hallucination by returning actual repo counts and top competitors.

Strengths
  • •Solves a genuine friction point: ChatGPT gives encouragement, not data, before 6-hour builds.
  • •Real API calls to 5 sources means output is verifiable and not LLM opinion.
  • •One-line install (uvx) and MCP integration means it plugs straight into Claude workflows.
Weaknesses
  • •Audience limited to developers; founders doing pitch decks get little value from technical saturation.
  • •Reality signal (0-100) is a single number; deeper analysis of *why* ideas fail would strengthen it.
Target Audience

Software developers, indie hackers, AI coding agents (Claude, Cursor)

Similar To

ValidatorAI · IdeaProof · Google Trends

Post Description

I kept wasting hours building things that already existed. Asked Claude to help me build a food delivery app — searched GitHub afterward, found 847 similar repos with 12 competitors over 1k stars. 6 hours wasted.

So I built an MCP server that does the searching before you write code. It scans 5 real-time sources (GitHub, Hacker News, npm, PyPI, Product Hunt) and returns a quantified reality_signal (0-100) with actual evidence — repo counts, star counts, top competitors, and pivot suggestions.

Example: "AI code review tool" → reality_signal: 90, 847 repos, top competitor reviewdog (9,094 stars), 254 HN mentions.

What it's NOT: not a business plan generator, not an LLM opinion wrapper. Every number comes from a real API call you can verify.

- Quick mode: GitHub + HN (default) - Deep mode: all 5 sources in parallel - Works with Claude Desktop, Cursor, Claude Code - One-line install: uvx idea-reality-mcp - Also available as web demo (no install needed)

GitHub: https://github.com/mnemox-ai/idea-reality-mcp Web demo: https://mnemox.ai/check MCP Registry: io.github.mnemox-ai/idea-reality-mcp

Built with Python + FastMCP, 120 tests, published on PyPI. Happy to answer any questions about the scoring algorithm or MCP integration.

Similar Projects