All Projects (6,181)

Autograph - a Mac app that installs Apple Mail signatures in one click logo
Autograph - a Mac app that installs Apple Mail signatures in one click

Design your signature on your Mac and install it into Apple Mail in one click. No copy-pasting HTML, no broken images, no subscription.

0
0
I built a fluid-sim latte art trainer over a weekend. I take on 2 MVP builds a month. logo
I built a fluid-sim latte art trainer over a weekend. I take on 2 MVP builds a month.

Learn latte art with your real pitcher. Follow the dot on your phone and pour the heart.

0
0
I built LayupLog, a free live basketball stats app focused only on one player. logo
I built LayupLog, a free live basketball stats app focused only on one player.

Built for parents, friends and scouts: log one player’s stats live from the stands. You tap the makes, misses and rebounds — LayupLog does the math.

0
0
Quotastrology - Because Codex resets are all based on vibes and so is astrology logo
Quotastrology - Because Codex resets are all based on vibes and so is astrology

Reset odds, banked-reset advice and a daily horoscope for your Codex usage limits, cast from every reset the Oracle has announced.

0
0
My side project is a multiplayer browser game set in real Italian cities logo
My side project is a multiplayer browser game set in real Italian cities

Passeggia tra vie, piazze, torri e monumenti di città italiane vere, ricostruite in un mondo pixel art, direttamente dal browser.

0
0
Built a whiskey tracker because I kept wondering if I overpaid for bottles. Signups are basically ze logo
Built a whiskey tracker because I kept wondering if I overpaid for bottles. Signups are basically ze

Track your whiskey collection and compare live prices across 26 US retailers. 1,080+ bottles. 21+ only. We do not sell alcohol.

0
0
Been building a music quiz game on the side for a few months logo
Been building a music quiz game on the side for a few months

Free online music quiz to play with friends: pick a playlist, invite your friends and name that tune in real time. No sign-up to play.

0
0
Made a game price site that shows what you'll actually pay at checkout, and tells you when Game Pass logo
Made a game price site that shows what you'll actually pay at checkout, and tells you when Game Pass

Compare game prices across every store and subscription. Real checkout prices, the cheapest way to play, all-time lows and what is leaving Game Pass.

0
0
I built OneMessage, a dating app where you get one message per person logo
I built OneMessage, a dating app where you get one message per person

A dating app where you get one message per person. No spam, no endless openers: they read it and decide.

0
0
e2e, super powerful, no accounts Discord-like chat app logo
e2e, super powerful, no accounts Discord-like chat app

Infinite anonymous chatrooms with auto deleting messages.

0
0
I wanted a calmer way to look at my Apple Health data logo
I wanted a calmer way to look at my Apple Health data

Akari brings your sleep, movement and food together on your iPhone and explains in plain words what they mean for your day.

0
0
Building an open-source MCP for AI verification. Looking for early testers. logo
Building an open-source MCP for AI verification. Looking for early testers.

Your AI passed the demo. Can it pass verification? Orinyx builds independent verification tools for developers testing healthcare AI agents.

0
0
I built a free map of realistic next jobs from where you are, with real pay for your state (US) logo
I built a free map of realistic next jobs from where you are, with real pay for your state (US)

Your career isn't a ladder. It's a map. The routes from where you stand: what they pay, how long they take, and how much of the work is digital.

0
0
I built a way to watch and fix my servers from my phone without opening a port logo
I built a way to watch and fix my servers from my phone without opening a port

Install the open-source cntrl agent and every server and device shows up in Console, live. It dials out, so there's no port to forward and no VPN to run.

0
0
Non-native English speaker here. I built a tool to understand hard sentences in papers without leavi logo
Non-native English speaker here. I built a tool to understand hard sentences in papers without leavi

Highlight the sentence you are stuck on. A short, clear explanation appears right on the page, in your language. No prompt to write.

0
0
I built a tool to automate media clipping system for newswires, PR agencies and communications teams logo
I built a tool to automate media clipping system for newswires, PR agencies and communications teams

News monitoring for the AI era: human web search, AI search and LLM visibility, in one client-ready report.

0
0
I built an app that publicly posts it when I code past my bedtime logo
I built an app that publicly posts it when I code past my bedtime

A small Windows and macOS app that watches your work apps after your cutoff. Close them in time and your streak grows. Leave one open and your fall goes on the public Blotter.

0
0
I turned my Android phone into a Windows trackpad. Did I waste my time? logo
I turned my Android phone into a Windows trackpad. Did I waste my time?

Turn your Android phone into a wireless trackpad and keyboard for your Windows PC. Open source, local network only, zero cloud.

0
0
I added Word documents and portable HTML pages to my foldable reader logo
I added Word documents and portable HTML pages to my foldable reader

We study allowing large language models (LLMs) to process arbitrarily long prompts through the lens of inference-time scaling. We propose Recursive Language Models (RLMs), a general inference paradigm that treats long prompts as part of an external environment and allows the LLM to programmatically examine, decompose, and recursively call itself over snippets of the prompt. We find that RLMs can successfully process inputs up to two orders of magnitude beyond model context windows and, even for shorter prompts, dramatically outperform the quality of vanilla frontier LLMs and common long-context and coding scaffolds (e.g., on GPT-5 by a median across the evaluated benchmarks of $26\%$ against compaction, $130\%$ against CodeAct with sub-calls, and $13\%$ against Claude Code) across four diverse long-context tasks while having comparable cost. At a small scale, we post-train the first model around the RLM. Our model, RLM-Qwen3-8B, outperforms the underlying Qwen3-8B model by $28.3\%$ on average and even approaches the quality of vanilla GPT-5 on three long-context tasks. Code is available at https://github.com/alexzhang13/rlm.

0
0
Open Co-Scientist: an open-source recreation of Google's AI Co-Scientist — looking for help and fe logo
Open Co-Scientist: an open-source recreation of Google's AI Co-Scientist — looking for help and fe

Generate, evaluate, and rank research hypotheses with a team of agents.

0
0