The biggest observation so far is that the latest models are night and day from LLMs from even 6 months ago (from OpenAI + Anthropic, Google is still very poor!)
https://github.com/timf34/habit-wallpaper
I built a habit tracker that actually works because it's impossible to ignore. Every day I get an email, click a link, check off my habits, and my laptop wallpaper updates with a GitHub-style heatmap showing my progress.Open source, you can sign up at habits.timfarrelly.com or self-host: https://github.com/timf34/habit-wallpaper
https://github.com/timf34/podscript
Built a tiny CLI that turns podcasts or YouTube videos into clean Markdown transcripts with speaker labels + timestamps.pip install podscriptUses ElevenLabs for high-quality diarization.https://github.com/timf34/podscript
I got tired of copy-pasting arXiv PDFs / HTML into LLMs and fighting references, TOCs, and token bloat. So I basically made gitingest.com but for arxiv papers. You can just append "2md" to any arxiv URL (with HTML support), and you'll be given a clean markdown version, and the ability to trim what you wish very easily (ie cut out references, or appendix, etc.)
https://github.com/timf34/pagesource
I wanted a way to download exactly what the browser loads for a webpage (i.e. all JS, CSS, images, fonts, etc.) with the original directory structure preserved (like what you see in DevTools → Network/Sources). Primarily for context/ prompting an LLM for creating similarly designed websites.Couldn't find a tool, so here it is:)
https://github.com/timf34/github-engagement-stats
GitHub only shows views and clones from the last 14 days — then deletes them. I wanted to see the longer-term impact of my open source projects, so I built a GitHub Action that saves daily snapshots and generates a simple Markdown report.It tracks stars, views, and clones across all your public repos — no setup, no dependencies, no secrets.
Focusing on Ireland for now, my home country. And admittedly I haven’t had a chance to work on this/ update it since September.
https://youtube-sort-by-likes.vercel.app/
Built this web app to find the best quality videos on a YouTube channel. Find it particularly useful for podcasts, long video channels, and finding music sets!YouTube API has really low quotas that seem to be a pain to increase, any way to make this work without using the youtube api would be greatly appreciated!Open sourced here: https://github.com/timf34/YouTubeSortByLikes
https://github.com/timf34/JSONDetective
Built this as I came across a really interesting JSON file, but it had 1.9 million lines and no public documentation! Point it at your JSON file, it identifies patterns and shows you the actual structure so you can easily understand it.
https://github.com/timf34/Article2Audio
Hey everyone,I made a web app that let's you pass the URL to an article, and receive back an MP3 narration of the article. With some configuring, you can also sync all mp3 files to a personal podcast (Spotify, Podbean, etc.) automatically.I've really enjoyed this for listening to longer articles that deserve to be an audiobook format!Work in progress but all the key functionality is there. Open...
https://github.com/timf34/PrintDirTree
A simple CLI tool that lets you print directory trees with options for custom exclusions. It's particularly useful for prompting ChatGPT.
https://github.com/timf34/Substack2Markdown
A Python script for converting Substack posts into Markdown. It supports both free and premium content, offering an easy way to archive and read these posts offline. Built with BeautifulSoup and Selenium, it’s adaptable for different browsers. Looking for feedback and open to contributions. Check it out on GitHub: https://github.com/timf34/Substack2Markdown