Zero-Code RSS Feeds with Mkfd: From Webpage to Feed in Minutes

Tim B — Mon, 10 Mar 2025 18:58:53 +0000

TBosak / mkfd

RSS feed builder created with Bun🥖 and Hono🔥- builds from webpages and/or REST API calls

🏃 Running locally

🍞 Bun installation ^{Visit bun.sh for more info}

curl https://bun.sh/install | bash

📦 To install dependencies

bun install

🚀 To run

bun run index.ts --passkey=your_passkey_here --cookieSecret=your_cookie_secret_here

➡️ Access the GUI at http://localhost:5000/

🐳 Running with Docker

🏠 Locally

docker build -t mkfd .
docker run -p 5000:5000 -v /local/mount/path:/configs -e PASSKEY=your_passkey -e COOKIE_SECRET=your_cookie_secret mkfd

📥 From Docker Hub

docker pull tbosk/mkfd:latest
docker run -p 5000:5000 -v /local/mount/path:/configs -e PASSKEY=your_passkey -e COOKIE_SECRET=your_cookie_secret tbosk/mkfd:latest

🔧 To Do

Locally testing subscriptions to feeds
Add all possible RSS fields to models
Add option for parallel iterators
Add form validation
Add selector suggestion engine
Add feed preview pane
Store/compare feed data to enable timestamping feed items
Create dockerfile
Create Helm chart files
Create GUI
Utilities
- HTML stripper
- Source URL wrapper for relative links
Amass contributors

View on GitHub

Mkfd is a zero-code RSS feed builder created with Bun and Hono, featuring an intuitive front-end that simplifies the entire feed creation process. Instead of manually writing scripts, you just open the interface, input a target URL along with CSS selectors, and Mkfd exports a YAML configuration file. Mkfd then uses that YAML to periodically build and update your feed. In this post, I'll show a high-level overview on how to use Mkfd’s user interface and show you how quickly you can start generating feeds from any webpage or REST API — without writing a single line of code.

🏃 Running locally

🍞 Bun installation ^{Visit bun.sh for more info}

curl https://bun.sh/install | bash

📦 To install dependencies

bun install

🚀 To run

bun run index.ts --passkey=your_passkey_here --cookieSecret=your_cookie_secret_here

➡️ Access the GUI at http://localhost:5000/

🐳 Running with Docker

🏠 Locally

docker build -t mkfd .
docker run -p 5000:5000 -v /local/mount/path:/configs -e PASSKEY=your_passkey -e COOKIE_SECRET=your_cookie_secret mkfd

📥 From Docker Hub

docker pull tbosk/mkfd:latest
docker run -p 5000:5000 -v /local/mount/path:/configs -e PASSKEY=your_passkey -e COOKIE_SECRET=your_cookie_secret tbosk/mkfd:latest

🧭 Navigating the Mkfd UI

Open http://localhost:5000 after running Mkfd. You’ll see a form that lets you input:

The Target URL of the webpage or API you want to convert into a feed.
A Schedule (in minutes) for how often Mkfd should re-scrape and rebuild the feed.

Mkfd provides a user-friendly interface (with SelectorGadget integration for figuring out the correct selectors) that lets you pinpoint elements on the webpage:

Article Container Selector – Identify the main wrapper for each item, e.g., .article or li.post-item.
Title, Link, Date Selectors – Extract the specific data fields for your feed.
Additional Fields – You can also include a description, author, or enclosures (such as audio, videos, or images).

The Aforementioned UI:

Note: If accessing remotely from outside localhost, the passkey used on deployment is needed to enter the UI but not to consume the feeds.

🗂️ Generating the YAML Configuration

After you specify your selectors and scheduling details, just hit Submit. Mkfd will create a YAML configuration file with a unique GUID — this file contains:

URL – The page or API endpoint you’re scraping.
Selectors – The custom CSS selectors for titles, links, dates, etc.
Schedule – The periodic interval (in minutes) for Mkfd to update your feed.

This YAML file lives within the project directory and serves as the master instruction set for the feed builder.

⚙️ How Mkfd Builds and Serves the Feed

Once your YAML is set up, Mkfd executes a periodic job to:

Scrape the Specified URL – Using your CSS selectors to retrieve the latest content.
Generate an RSS Feed – Formatting the scraped data into a valid feed structure.
Expose the Feed – Mkfd will automatically serve the feed from an endpoint for the end user to consume.

💡 Tips and Tricks

Selector Fine-Tuning: If your feed items aren’t showing as expected, double-check your CSS selectors.
Multiple Feeds: You can create multiple feeds for different websites or data sources. Each feed is built using its own web worker.
Scheduling Intervals: Avoid overly frequent scrapes (e.g., every minute) to reduce server load and potential IP blocking. A common interval is 15 minutes to 1 hour.
Version Control Your YAML: If you’re setting up multiple feeds or complex data extractions, keep your YAML in version control for easy rollback or iteration.

✅ Conclusion

Mkfd offers a straightforward, no-code solution for generating custom RSS feeds out of any webpage or REST API.

Scrape Data using simple CSS selectors—no coding required.
Generate a YAML file that captures your scraping logic and schedule.
Automatically Build and Serve the feed on a configurable schedule.

Whether you’re a seasoned developer or just need a simple feed builder, Mkfd’s browser-based UI, integrated selector tools, and one-click YAML generation make it easy to convert the web into feeds you control — all without writing a single line of code.

Happy Building! Check out the GitHub repo for more details and/or to contribute to Mkfd’s development.

Janus - Visualize Your Journey

Tim B — Sun, 12 Jan 2025 03:41:33 +0000

This is a submission for the GitHub Copilot Challenge : Transitions and Transformations

What I Built

Janus is a tool for visualizing transitions and transformations via interactive timelines. It is built using Angular and served via Bun.

Demo

https://janus.baraniservices.us/

Repo

TBosak / janus

Janus is a tool for visualizing transitions and transformations via interactive timelines

Janus

🏃 Running locally

🍞 Bun installation ^{Visit bun.sh for more info}

curl https://bun.sh/install | bash

📦 To install dependencies

bun install

🚀 To run

bun run start

➡️ Access the GUI at http://localhost:6200/

🐳 Running with Docker

🏠 Locally

docker build -t janus .
docker run -p 6200:6200 -v /local/mount/path:/uploads -v /local/mount/path:/timelines janus

View on GitHub

Copilot Experience

Throughout the process, I used Copilot for code completion and chat. In chat, I tested out several different models - Claude 3.5, GPT 4o, o1, and o1-mini. I preferred o1 most of all, as I'm most familiar with it through my experience using ChatGPT. I periodically corrected syntax during the challenge using Copilot, occasionally used it for debugging, and used GPT o1 a lot for more organization and high-level questions. I find that Copilot keeps me more in a flow state, as I don't stray from my IDE.

Conclusion

Janus was an exciting project that allowed me to combine the power of Angular and Bun to create an interactive and visually engaging tool. Participating in the GitHub Copilot Challenge gave me the opportunity to explore how Copilot can enhance the development process by streamlining code completion, debugging, and overall workflow. By leveraging Copilot alongside other tools, I was able to stay focused and maintain a productive rhythm while building Janus. I'm thrilled to have contributed to this challenge and to share Janus with the developer community. Thank you for the opportunity!

DEV Community: Tim B

[Boost]