Today, during the day, I received a request from a friend who sometimes uses the search function on Ecency to find specific content on the Hive blockchain. He uses it mainly to curate posts more effectively. What he really needs is a way to search for certain words that were written inside the content of a post, not just in the title or tags.

He told me that there used to be a Discord bot that scanned the newest blog posts and sent a private message if it found something interesting. Sadly, that bot is no longer available.

That reminded me of a tool I had developed some time ago. It was built using hafSQL from Hive account@mahdiyari and allowed users to search for posts made by specific Hive accounts. The team at Hive account@hivewatchers also found the tool helpful, especially for looking up user behavior over time.

So I took the same codebase, adjusted it a bit, and built something new. Well, I should say I let it build itself, with a few clever prompts and some help from my AI assistant. The new version of the tool can now search up to ten keywords inside posts from the past three days. It uses hafSQL to scan the post bodies, not just the titles, and returns any matches it finds.

Right now, we do not really have a great notification or deep content search function on Hive that works like this, so I think this can be very useful. Especially for curators who want to quickly spot trending topics or specific terms that are being discussed in real time.

I have now also uploaded the tool to Render.com, where it can be accessed at the following URL:

https://keywordsearch.vercel.app

---

If the frontend or the tool is not reachable at some point, it does not mean that the whole service is broken or down. The reason is actually quite simple. I am currently using the free plan from Render to host the app. And with that plan, the service automatically goes to sleep after some time when nobody is using it.

So if you try to access the tool and it seems offline, just be patient. Click the link, wait a few moments, and the service will start up again by itself. This is completely normal for apps running on a free Render instance.

When it is waking up, it might take around 30 to 60 seconds before everything loads properly. But once it is online again, everything should work just fine. On your end, you might just see a loading screen or nothing happening at first — but do not worry, that is just the app starting up in the background.

Readme File:

🐝 Hive Keyword Tracker

A powerful, real-time search tool for the Hive Blockchain. This application allows users to monitor specific keywords across all Hive posts created in the last 3 days using a direct SQL connection to a HAF (Hive Application Framework) node.

✨ Features

• Real-time HAF SQL: Queries the blockchain state directly for instant results.
• Smart Keyword Monitoring: Track up to 10 keywords simultaneously.
• Time Window: Automatically scans the last 72 hours (3 days) of blockchain activity.
• Rich Post Previews:
  • Extracts and displays the first image from Markdown or HTML.
  • Highlights keyword context in snippets.
  • Shows author, date, and category/tags.
• Platform Choice: Open posts directly in your favorite interface:
  • PeakD
  • Ecency
  • Hive.blog
• Theme System: Built-in Dark Mode and Light Mode with persistence.
• Local Persistence: Remembers your keywords, platform preference, and theme settings.
• Debug Console: View the exact SQL queries generated and execution logs.

🏗 Architecture

Since web browsers cannot connect directly to PostgreSQL databases (due to security protocols), this project uses a lightweight Node.js Middleware architecture:

1. Frontend (React + Vite): User interface and settings.
2. Middleware (Express): Receives requests from the frontend, securely connects to the HAF SQL database, and returns sanitized JSON data.
3. Database (HAF SQL): The public Hive Application Framework node (provided by Mahdiyari).

🚀 Getting Started (Local Development)

To run this project locally, you need to run both the Backend (Middleware) and the Frontend (React).

Prerequisites

• Node.js (v18 or higher)
• npm

1. Install Dependencies

npm install

2. Start the Backend Server

Open your first terminal window and run:

node server.js

This starts the middleware on http://localhost:3000.

3. Start the Frontend

Open a second terminal window and run:

npm run dev

This starts the UI (usually on http://localhost:3001).

4. Configure Connection (optional)

1. Open the App in your browser.
2. Go to Settings.
3. Ensure Middleware URL is set to: http://localhost:3000/search.

---

☁️ Deployment (Render.com)

This project is optimized for deployment on Render as a "Web Service".

Build Command

Render needs to build the frontend and install backend dependencies:

npm install && npm run build

Start Command

Render needs to start the Node.js server (which serves the built frontend):

node server.js

Environment Variables

No manual environment variables are required. The application automatically detects:

1. The PORT provided by Render.
2. The environment (Production vs. Local) to adjust the API endpoint automatically to /search (relative path).

---

🛠️ Tech Stack

• Frontend: React 18, TypeScript, Tailwind CSS, Lucide React
• Backend: Node.js, Express.js, pg (PostgreSQL client)
• Database: HAF (Hive Application Framework) PostgreSQL
• Tooling: Vite

⚠️ Troubleshooting

"Failed to fetch" / Connection Error

1. Local: Ensure node server.js is running in a separate terminal. Check your Settings > Middleware URL is http://localhost:3000/search.
2. Production: Check the deployment logs. Ensure the build command successfully created the dist/ folder.

Database Errors (Status 500)

• Check the Debug console at the bottom of the app.
• If the HAF node is down or the query times out, the error details will appear there.

❤️ Credits

• Creator: @louis88
• HAF Node: Public endpoint provided by Mahdiyari.

---

Thank you very much! If you have any requests for changes, questions, or suggestions, please feel free to write them in the comments or tag me. Until then... have a wonderful Advent season!