myGeeKy

See

See what is moving in your field

“I read the twenty sites, so you read one.”

The papers behind the code, and the people behind the papers

Matched to your own research topics, not a generic popularity list.

Papers in your field gaining citations fastest, measured in citations per month
New tools whose code is on GitHub. Click a paper to see its repo and the people working on it
A green P on any repo that has a published paper
Your field's repos as a daily market: stars, commits and downloads, and what rose and fell since yesterday
Models trending on Hugging Face in your field

It reads OpenAlex, arXiv, bioRxiv, Hacker News, PyPI, Hugging Face, Product Hunt, ORCID, Google Scholar and GitHub. GitHub is one of them, not the point.

Research view: papers rising in your field

What's new in your field, not everyone's

Fresh items from arXiv, bioRxiv and Hacker News, kept only when they genuinely match your interests: two of your terms, or one exact phrase.

Each item says why it's there ("matches gwas, fine mapping")
🔭 items from outside your bubble are marked, on purpose

And see what myGeeKy knows about you

The Model tab draws it as a living map: you in the middle, your research field around you, what you've been into lately pulsing further out, and today's new territory on the edge. Correct it whenever it's wrong.

Clicks teach it a little, stars and forks more, follows the most
Interests fade with a 30-day half-life, so it follows you as you change
🔭 20% always stays open for new territory, so you never end up in a bubble
A separate model learns who follows you back

your fieldlately intonew territory

● Keywords that mean something

Add "AI" and myGeeKy also looks for LLMs, deep learning and AI agents. Add "GWAS" and it brings in genomics and statistical genetics.

● Your own words still work

Words not in the dictionary work as plain words, and are passed on, through your beacon or with one click, so the next dictionary learns them.

🔁 It keeps learning

A new dictionary reaches everyone by itself, at most once a week. No server, nothing hidden: the model is a few hundred lines you can read.

Learn

Learn what comes next, with a road map to get there

“You haven't used Nextflow yet, and five repos near your work just did. Shall we start?”

Any skill becomes a road map

Rising Skills shows what's growing in your field and missing from you. A protocol turns any skill into steps: R basics → statistics with R → a real project to rerun → your own work.

A short check skips what you already know: your own repos as evidence, a quiz, or "I can do these"
Each step comes with repos, papers, free books and 📘 guides from a project's own documentation
14 reviewed protocols: Git, algorithms, a first web app, machine learning, a first pull request · R, Python, GWAS, Nextflow, single-cell RNA-seq · Excel to R, statistics for experiments, figures for papers, images into numbers with Fiji
Any other skill: built from GitHub, papers and protocol journals by a small model (about 400 KB) on your device
📚 Free courses, books and tutorials, picked by hand; 🎉 wins: new stars, forks and citations of your work

Ten questions, and not one of them invented

Every protocol has a question bank. An attempt draws ten at random: half basics, then applied, and a few of the mistakes people really make.

Each question was written from one real source and quotes the sentence that proves the answer, and shows you that source once you've answered
Code questions are run, and sums recomputed, before a question ships
Only your best score counts, one attempt a day, and 🚩 drops a question you think is wrong
Nothing about you goes online unless you ask. Your name and your scores are two separate choices, agreed by writing accept on a GitHub issue from your own account; withdraw undoes it
Low scores are never shown anywhere: the top three of the month and the first people to finish a protocol are all the page ever lists

A quiz question, answered: why that answer is right, and the source it came from

Meet

Meet the people working on the same problems

“These four work on what you work on, and they tend to follow back.”

Friends, not followers

Researchers who share your work and are likely to follow you back, plus domain experts who rarely follow anyone. Celebrities, mass-followers and dormant accounts are skipped.

Live: what your circle pushed in the last hour, one card per person
Repos worth your first pull request: starter issues, and maintainers who actually merge outside work
Collaborators whose skills complement yours
You follow, star and message by hand, always. myGeeKy introduces; it never acts for you

Live tab: what the people near your work did in the last hour

Signals: gratitude, not small talk

A quiet way to tell someone their work mattered to you: no server, no text, nothing to moderate, and nothing to answer.

Private: each signal is encrypted so only its recipient can read it. Nobody can see who you signalled, or who didn't answer
No reply needed: every signal says so, and you never see "unanswered"
🤝 only if you both want it: "let's collaborate" stays hidden until it's mutual
Courteous by design: once a month per person, a few new people a week, a "not taking signals" switch, mute and block

Catch

Catch the opportunities while they are still open

“Two deadlines this week. Both still open; I checked the page myself.”

Verified jobs, funding and conferences

PhD places, postdocs, fellowships, travel grants and meetings in your field, and internships, hackathons and software jobs. Each one is checked against its own page before you see it.

Sorted by deadline, and gone the moment it passes: never a list of things you're already too late for
A listing in Arabic, French, German, Spanish or Chinese is read, translated and still checked, then shown in English
Matched to you on your own device: nobody learns what you looked at
👍 / 👎 teach your model; 🚩 tells the maker one looks fake

🗓 Seasons, for the ones that come back every year

A scholarship doesn't stop being real when it closes: it returns next April, and the preparing takes longer than the applying.

“Expected in about 12 weeks”, while there is still time to get transcripts certified and a proposal written
It tells you how many years of evidence stand behind that expectation
It never pretends to know next year's date: a window read on a page says open now, anything ahead says expected

Seasons: a scholarship that comes back every year, with the weeks of preparation it needs

A geek in a hoodie working on a laptop against a brick wall, a myGeeKy sticker on the lid

Who it's for

For anyone who works alone

The student in a lab where nobody else writes code. The researcher in a city with nobody in their field. You might be the only one in your building who does this; you are not the only one in the world.

🔕 No noiseNo feed, no likes, no streaks, no notifications.
👋 Friends, not followersPeople near your work, and who's likely to answer.
🚪 Nobody sees you lookingNo account, no sign-up, no telemetry.
🌙 It waitsClose it and nothing keeps running. It never pings you.

💻

You mostly write code

Repos that welcome your pull requests, people building what you build, the tools rising in your corner of software, and interviews, internships and hackathons under 🎯 Chances.

Start with: Git and GitHub for working with others

📊

You analyse your own experiments

No repos needed: it reads your CV and your papers. The gentlest protocols start where you are, a spreadsheet of results, and end with a figure and a test you can defend.

Start with: From Excel to R, with your own data

🧬

You analyse data

Papers gaining citations fastest with their code, pipelines and methods rising in your field, and the postdocs, fellowships and conferences that fit your topics.

Start with: R for data analysis

You say which fits you best when you set it up, and change it whenever you like; your model learns from what you actually open. Code, data, statistics, figures and pipelines, never laboratory protocols or bench instructions: it can't review bench safety or know your lab's rules, so it stays out of it.

On your phone

Your world, in orbit

You in the middle, your circle on the inner ring, the world on the outer one. Spin a ring with your thumb, and the planet that locks into the lens opens above it.

Live, People, News, Papers, Repos, AI models, 🌱 Grow and 🎯 Chances, each a planet that shows how much it holds
Tap something to read and the rings fold away into a slim dock
Set up in one scan (⚙ → Connect your phone on the desktop panel), or sign in with GitHub
For Android, coming to Google Play soon

The myGeeKy phone app: the orbit with Chances, a protocol's road map, and conferences

Privacy & trust

What it will never do

“I keep all of this on your computer. I never post for you, and I never tell anyone you were here.” Read the privacy policy.

Never acts as you.

It never follows, stars, forks, posts or comments: there is no code for it anywhere in the project.

Never leaves your machine.

The model that ranks it all runs on your computer. No server, no account, no telemetry.

Never holds your keys.

Tokens live in your system's keyring, checked for least privilege, never logged or printed.

Never takes your data.

It stays on your computer, plus a private repo you own, if you want sync.

Never sends anything behind your back.

If it breaks, you read the whole report first, personal details stripped, and you decide.

🔍 Token 1 · read-only

Reads public profiles, repos and followers.

Repositories: Public repositories
Permissions: Followers: Read-only
Refused if: it can write to any repo, can follow people, or is your Signals token

✍️ Token 2 · Signals (optional)

Writes your mygeeky-beacon repo, and nothing else.

Repositories: Only mygeeky-beacon
Permissions: Contents: Read and write
Refused if: it can't write to the beacon, or is token 1 again

Setup checks what each token can actually do, and refuses one that doesn't fit its job. A GitHub account is optional: it unlocks who'd follow you back, your activity and Signals; everything else works without one. MIT-licensed and open to read: every promise here is a file you can check.

Get it

Up and running in two minutes

🪟 Windows

Download MyGeeKySetup.exe
Next → Next → Finish

Nothing needs installing first: it sets up Python and Git for you, no admin rights. If Windows says “Windows protected your PC” (not code-signed yet), click More info → Run anyway.

🐧 Linux · macOS · anywhere

curl -fsSL https://mygeeky.org/install.sh | sh

or with pip:

pip install mygeeky
mygeeky setup

Run the one-liner again to update. Macs get myGeeKy.app in Launchpad and Spotlight.

Then just type what you work on. That's the whole setup: the panel opens, and the first run sets up its own window by itself.

Code signing policy

How Windows releases are signed

Windows releases of myGeeKy will be signed through the SignPath Foundation's free code signing program for open-source projects (application pending). Until approved, releases are unsigned.

Only binaries built by this project's public GitHub Actions workflow from the source in github.com/mygeeky/myGeeKy will be signed.
Committers and reviewers: Alsamman M. Alsamman
Approvers: Alsamman M. Alsamman
Privacy: this program will not transfer any information to other networked systems unless specifically requested by the user or the person installing or operating it. Its job is to read public data from GitHub and research services for the account you set up, and to check PyPI for updates once a day; it sends no personal data or telemetry. See the privacy policy.

Your geeky friend who helps you grow, quietly.

Lives quietly on the edge of your screen