Forem: goldct

I Built an AI Fortune Telling App with SwiftUI — Meet Arcana Stars

goldct — Sun, 19 Apr 2026 00:08:34 +0000

Arcana Stars: AI-Powered Fortune Telling

What happens when you combine Tarot, astrology, and AI?

What It Does

78 Tarot Cards with AI readings
12 Zodiac Daily Horoscopes
Crystal Ball — AI yes/no predictions
Love Compatibility — zodiac matching

Tech Stack

SwiftUI (zero UIKit)
GLM-4-Flash API (free tier!)
StoreKit 2 subscriptions
Zero third-party dependencies

Why

I wanted to prove a solo dev can build a polished AI app without expensive APIs or a design team.

Status

Currently under App Store review! 🤞

Follow for updates. Happy to share technical details if interested!

Built with ❤️ and AI.

From Zero to MCP: How I Automated My Entire Workflow

goldct — Tue, 14 Apr 2026 11:56:07 +0000

Before MCP

Every task was manual:

Write SQL, copy results, paste in chat
Open files one by one for AI to read
Manually create GitHub issues
Check servers in terminal

After MCP

Everything happens in one chat:

Hey Claude, what are the top selling products this month?
Read the config file and tell me what needs fixing
Create an issue for this bug
Check if the container is running

The Setup

I installed 5 MCP servers in 10 minutes:

PostgreSQL — database queries
Filesystem — file access
GitHub — repo management
Docker — container status
Brave Search — web lookup

Finding Tools

Browse 900+ at mcphello.com with quality scores and one-click install.

The Result

I went from 2 hours of context switching per day to near zero. MCP is a game changer.

What would you automate with MCP? Share your use case!

Why Every Developer Should Care About Model Context Protocol

goldct — Tue, 14 Apr 2026 11:50:45 +0000

The Shift

AI assistants are no longer just chatbots. They can now access your database, read your code, manage your servers, and browse the web.

The key enabler? Model Context Protocol (MCP).

What MCP Does

MCP gives AI a standard way to connect to external tools. One protocol, many integrations:

Database access (PostgreSQL, MySQL, SQLite)
File operations (read, write, manage)
Web interaction (search, scrape)
DevOps (Docker, Kubernetes, CI/CD)
Communication (Slack, Discord, email)

Why Developers Should Care

10x productivity — AI handles the boring stuff
No more context switching — everything in one chat
Open standard — not locked to any vendor
Growing ecosystem — 900+ tools already

Getting Started

Browse tools at mcphello.com — 900+ MCP tools with quality scores and one-click install.

The future of development is AI-assisted. MCP makes it possible.

The Developer Guide to MCP Quality: How to Evaluate MCP Servers

goldct — Tue, 14 Apr 2026 11:43:41 +0000

Not all MCP servers are equal. Here is how to evaluate them.

Criteria That Matter

1. Code Quality

Does it have tests?
Is the code well-structured?
Does it follow MCP conventions?

2. Documentation

Clear README with examples?
Configuration guide?
Troubleshooting section?

3. Maintenance

Recent commits?
Responsive to issues?
Active contributors?

4. Community

GitHub stars and forks
User reviews
Community contributions

The Easy Way

At mcphello.com, every tool gets a 0-100 quality score based on all these factors. No need to evaluate manually.

900+ tools scored
Updated every 6 hours
One-click install commands
Free, no signup

What criteria matter most to you when choosing an MCP server?

10 MCP Servers That Will Save You Hours Every Week

goldct — Tue, 14 Apr 2026 11:38:23 +0000

Time is money. Here are 10 MCP servers that will make you more productive:

1. Filesystem

Read and write files from your AI assistant. Essential.

2. PostgreSQL

Query databases directly from chat. No more copy-pasting.

3. GitHub

Manage repos, issues, PRs without leaving your AI tool.

4. Puppeteer

Automate browsers with AI. Testing and scraping made easy.

5. Google Drive

Access your docs and sheets from Claude.

6. Slack

Send messages, read channels, manage workflows.

7. Memory

Persistent memory for AI across conversations.

8. Brave Search

Web search directly from your AI assistant.

9. SQLite

Lightweight database access. Perfect for local projects.

10. Docker

Manage containers from chat. DevOps without the CLI.

Discover 900+ more at mcphello.com

Which MCP server saves you the most time? Comment below!

How Python Agents Helped Me Catalog 900+ AI Tools

goldct — Tue, 14 Apr 2026 11:33:05 +0000

The Challenge

Keeping track of 900+ MCP servers manually? Impossible. I needed automation.

My Python Agent Architecture

Collector Agent

Crawls GitHub API every 6 hours looking for new MCP repos.

Analyzer Agent

Scores each tool on quality, docs, activity.

Writer Agent

Generates review articles for top tools.

Ops Brain

Coordinates everything. Tells me what to focus on.

Results

900+ tools cataloged automatically
Quality scores for every tool
Deep reviews generated weekly
Zero manual data entry

The Tech Stack

Python 3.12 with asyncio
GitHub REST API + GraphQL
Playwright for social media posting
Cron for scheduling

Try It

See the results at mcphello.com

Questions about the agent setup? Ask in comments!

The Rise of MCP: Why AI Tool Integration is Exploding

goldct — Tue, 14 Apr 2026 11:26:58 +0000

MCP is Growing Fast

Model Context Protocol went from zero to 900+ servers in months. Here is why.

The Catalyst

Anthropic open-sourced MCP for Claude. Then Cursor, Windsurf, and others adopted it. Now it is becoming the standard for AI tool integration.

The Numbers

900+ servers on GitHub
Multiple categories: databases, files, web, DevOps, creative
Growing daily: new servers every week
Directory: mcphello.com tracks them all

What This Means

AI assistants are no longer isolated. They can:

Query your database
Read and write files
Browse the web
Manage your GitHub
And much more

Getting Started

Get an MCP-compatible AI tool (Claude, Cursor, etc)
Browse tools at mcphello.com
Install and configure
Done!

The MCP ecosystem is just getting started. Exciting times ahead.

Open Source Projects I Found While Building an MCP Tool Directory

goldct — Tue, 14 Apr 2026 11:20:07 +0000

While cataloging 900+ MCP servers for mcphello.com, I discovered some incredible open source projects. Here are my top finds:

1. Everything MCP

A meta-server that bundles common MCP tools. Great for getting started quickly.

2. Database Connectors

PostgreSQL, MySQL, SQLite, MongoDB — all have MCP servers now. Query your data directly from Claude.

3. Browser Automation

Playwright and Puppeteer MCP servers let AI control browsers. Powerful for testing and scraping.

4. File System Tools

Safe file read/write for AI agents. Essential for any coding workflow.

5. Creative Tools

Image generation, music, video — yes, MCP servers for creative work exist too.

Browse all 900+ tools at mcphello.com with quality scores, install commands, and reviews.

What hidden gem MCP servers have you found?

How to Set Up Your First MCP Server in 5 Minutes

goldct — Tue, 14 Apr 2026 11:14:30 +0000

What You Need

Claude Desktop (or Cursor)
5 minutes

Step 1: Find a Tool

Go to mcphello.com and browse 900+ MCP tools. Pick one.

Step 2: Copy Install Command

Every tool on mcphello has a one-click install command. Copy it.

# Example: filesystem MCP
npx @anthropic/mcp-filesystem

Step 3: Add to Config

Open your Claude Desktop config:

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["@anthropic/mcp-filesystem"]
    }
  }
}

Step 4: Restart

Restart Claude Desktop. Your MCP server is live!

Step 5: Use It

Ask Claude to read files, write data, query databases — whatever your MCP server supports.

Want More?

Browse 900+ tools at mcphello.com:

Quality scores for every tool
Category filters
Search by name or function
Updated every 6 hours

What MCP server did you set up first?

MCP vs Plugins vs Extensions: Which AI Integration Should You Use?

goldct — Tue, 14 Apr 2026 11:09:02 +0000

AI tools need to connect to external data and tools. But how? There are three main approaches:

1. MCP (Model Context Protocol)

Standard, open, growing fast

Created by Anthropic for Claude
Now supported by Cursor, Windsurf, and others
Standardized JSON-based protocol
Growing ecosystem: 900+ tools

2. Plugins (ChatGPT-style)

Proprietary, platform-locked

Only works on one platform
Requires approval process
Limited to specific AI products

3. Custom Extensions (VS Code-style)

Flexible but complex

Full control but lots of work
No standard protocol
Hard to reuse across tools

My Recommendation

If you are building AI integrations in 2025, MCP is the way to go. It is open, standardized, and the ecosystem is exploding.

Browse 900+ MCP tools at mcphello.com — quality scored, one-click install, free.

What integration approach do you prefer?

Building a Side Project with 7 AI Agents: Lessons Learned

goldct — Tue, 14 Apr 2026 11:03:38 +0000

The Project

mcphello.com — an MCP tool directory with 900+ tools. Built and maintained entirely by AI agents.

The Agent Team

Collector — finds new MCP repos on GitHub
Analyzer — scores each tool
Writer — generates review articles
Publisher — updates the website
Ops Brain — coordinates and strategizes
Social Agent — publishes to social media
Dev.to Agent — writes and publishes articles

Lessons

What Worked

Clear agent responsibilities (single purpose each)
JSON-based data sharing between agents
Cron-based orchestration (simple and reliable)

What Did Not Work

Complex agent-to-agent communication (over-engineered)
LLM-based orchestration (unreliable, expensive)
Real-time agent networks (unnecessary for this use case)

The Stack

Next.js 16 + Vercel (frontend)
Python agents (backend automation)
GitHub API (data source)
Playwright (social media publishing)

Result

A fully automated directory that updates itself. Zero manual work after setup.

Check it out: mcphello.com

Any questions about the agent architecture? Happy to share details!

I Scored 900+ MCP Servers So You Do Not Have To

goldct — Tue, 14 Apr 2026 10:58:13 +0000

Quality Matters

Not all MCP servers are created equal. Some have amazing docs and active communities. Others are abandoned side projects with broken installs.

I built a scoring system to separate signal from noise.

How Scoring Works

Each MCP server gets a 0-100 score based on:

GitHub Stars (25%) — Community popularity
Documentation (25%) — README quality, examples, guides
Recent Activity (25%) — Commits in last 30 days
Issue Response (25%) — How fast maintainers respond

Top Scorers

Check the full list at mcphello.com

The Tool

mcphello.com catalogs 900+ MCP tools with:

Quality scores for every tool
One-click install commands
19 categories
Updated every 6 hours

Free. No signup. Just search and install.

Built with Next.js + Python agents. Feedback welcome!