Everything I publish, in the order I published it: dispatches from Twitter, videos from YouTube and notes from my vault, archived here in full — text, images, video and links — so they outlive the platforms.
A native macOS app for playing with LangChain that ships with a self-contained, embedded, Apple-notarisable Python 3.11 and up-to-date LangChain libraries
This project is a ‘batteries-included’ macOS UI for LangChain. It ships with a self-contained, embedded, Apple-notarisable Python 3.11 and LangChain libraries. It also includes documentation on how to update langchain dependencies over time.
Why? Access to LangChain should be easy for those who want to change the world.
I saw my girlfriend fully embrace GPT-4 via an app I built for her. Even though it’s “just a UI to GPT-4”, it comes batteries-included. This allowed her to fully focus only on her projects and not the technology.
The hard, and tedious, part - like installing brew, getting the right Python version, creating the right venv, updating langchain, choosing and learning an IDE - should not a barrier to entry for her.
UI Evolution
Here you can see the small user-facing increments.
The first (working) UI!
Got LangChain wired up!
Development Notes
To build with XCode, please create APIConfig.xcconfig under $ROOT/EmbeddedPythonSpike/Backend with this content
@upcycledwords very cool! might be fun to disallow backspace. this would truly make it a 'stream' of consciousness
@aashmngo@upcycledwords Agreed! No backspace and no return-key changes the experience a lot. It has helped me learn to love mistakes and not worry about typing accuracy at all. -@upcycledwords!–My typos make little fluffy clouds now. Feel free to give it a spin: getstreamline.app is open-source!
@kepano@obsdmd@upcycledwords Nice one! I wound up building a stream-of-consciousness app for my Obsidian (getstreamline.app) because my version of a plugin still distracted me. Will try yours!
JamGPT gets you to ChatGPT the instant you need it — one keystroke from whatever app you’re in, with no tab-hunting or logging back in. I kept my go-to prompts a click away for the things I do over and over, like translating or tightening up a sentence. It’s free to use, and ready the moment you download it.
Which galaxy do you like best? Look at the activity of ERC20 tokens as a map of galaxies in the Alexis pudding project. t.co/nHp0LUYyE3
Powered by #EthplorerAPI #ERC20 t.co/tLjpRIfTVq
I wanted a calm, blank space to get my thoughts out before the day’s noise set in — no menus, no formatting, no friction. As you write, Streamline surfaces your existing notes so a morning brain-dump connects straight into the system you already keep. Everything stays on your machine, with no account and nothing tracked.
Another open-source OSINT tool I made, this one for Ethereum token ecosystems. Pick any token and instantly see where the money flows — who’s moving it, to whom, and how much. The biggest players and tightest relationships rise to the top, so you can filter out the noise, zero in on what matters, and pull a real story out of the chain. Export what you find to share it. Try it.
Open-Source Intelligence (OSINT): connect the dots, follow the money, find the story in Ethereum ERC-20 ecosystems — pulls holders and transfers from public chain data, lays them out as an interactive graph and lets Louvain community detection surface the clusters worth asking about. Runs in the browser, no account needed.
@vanilla_chief Mega Idee – das von Harry Potter sieht auch richtig gut aus als Illustration. Danke! Ich melde mich gerne, wenn das als Feature dann fertig ist :)
Machine learning: a macOS meeting-buddy for those (like myself) who talk too much, or too little, when they get nervous — an on-device sound classifier answers "is someone talking?" every few milliseconds, tracks your airtime and shows a gentle traffic light. No recording, no cloud, no sign-up.
24 stars · 2 forks · SCSS
Unblah – For those (like myself) who talk too much, or too little, when they get nervous in meetings
Unblah is a little “meeting-buddy”-app that helps…
Keep track of how long you’ve been speaking,
Catch yourself before you talk too much, or too little