The free tool store you have been walking past

On the internet, there exists a place filled with thousands of genuinely helpful programs that cost nothing because the people who made them chose to share them. Some were created in newsrooms. You have almost certainly never visited it, since the whole site looks as though it was made for programmers.

Its name is GitHub, and this week I want to show you how to walk through the front door.

You do not need to learn programming. You do not need to know anything about code. What you do need are four things: an explanation of what it is, a way to distinguish a solid tool from a neglected one, the steps for downloading one, and the first three worth trying.

What GitHub actually is

Imagine a hardware store where every tool is free, and the builder has left the blueprints beside it on the shelf.

That is GitHub. It is a place where people keep the software they are developing, and most of them allow anyone to view it, download it, and use it. Large companies host projects there. So do solo developers working evenings and weekends. So do newsrooms such as The Marshall Project and Stanford's Big Local News, which create reporting tools and then give them away at no cost.

Programmers use it to collaborate on code. You can use it like a hardware store: go in, locate the tool, and take it home.

The one word you need: repository

Each project on GitHub sits in its own folder, and that folder is called a repository, or "repo" for short. One repo, one tool.

When someone says "check out this repo," they mean "go look at this project's folder." That is the entire meaning. Inside, you will see the program's files, some instructions, and a record of every change anyone has ever made.

How to tell a good tool from a dead one

This is the main skill, and once you know where to look it takes roughly fifteen seconds. When you open a repo page, inspect three things.

1. The star count, top right. A star is a click that means "this is useful." It is similar to the number of reviews on a product page. Nobody stars something that does not work.

Roughly speaking:

• Under 50 stars: very new, very specialized, or hardly used at all. Proceed carefully.
• A few hundred: real, with an audience.
• Thousands: well known and broadly trusted.
• Tens of thousands: major software.

2. The date of the latest update. Many people overlook this, but it matters more than stars. Look near the file list for wording such as "last commit 3 days ago" or "updated 2 years ago." A "commit" is simply a saved change.

If the newest change is from this year, someone is still maintaining it. If the last change was four years ago, it has been abandoned. It may still function, but no one is repairing it when it fails.

I checked every tool below against this standard before recommending it, and that removed several popular options from my list.

3. The README. Scroll downward. Beneath the file list is a long explanation called the README. That is the instruction manual: what the tool does, who it is for, and how to install it. If the README is clear and includes pictures, the maker cares whether ordinary people can use it. If there is no README at all, move on.

How to search for something useful

Use GitHub's search box, but phrase your search the way you would explain the problem, not the way a programmer would.

Search for "extract table from PDF" rather than "PDF parser." Search for "transcribe audio" instead of "speech to text ASR." Then, and this is the key step, open the Sort dropdown on the results page and select Most stars. With that single click, the proven tools rise to the top and the abandoned experiments sink down.

Honestly, the easier path is this: ask ChatGPT or Claude, "what is a free open source tool for [your problem], and is it still being maintained?" Then take the name it gives you and search for it on GitHub yourself using the three checks above. Let the AI identify possibilities. You confirm them.

How to actually download one

This is the part that scares people away, though it should not. There are two approaches, and the first one is completely harmless.

The easy way: Releases

Many tools come packaged as a standard installer, just like any program you have installed before.

On the repo page, look on the right side for a heading labeled Releases. Click it. You will see a list of versions, with the newest first. Beneath the newest version is a section called Assets containing files to download. Choose the one that matches your machine: something with "win" or ".exe" for a PC, something with "mac" or ".dmg" for a Mac.

Download it. Double click it. Finished. That is all there is to it.

The other way: Download ZIP

If there is no Releases section, find the green Code button near the top. Click it, then choose Download ZIP. This gives you a zip file containing everything in the project. Unzip it and read the README to learn what happens next.

That second route is more likely to require technical assistance. If you reach that point and it starts asking you to open a terminal, that is your cue to hand it to whoever manages your tech, or to skip it.

Three to try this week

I verified all three directly on GitHub before writing this.

1. Tabula: pull tables out of PDFs

Your county sends the budget as a PDF. The state sends salary schedules as a PDF. Election results, assessment rolls, everything in PDF form, and none of the numbers will copy and paste without turning into nonsense.

Tabula is a free program for Mac and Windows. Open it, load the PDF, drag a box around the table with your mouse, and it produces a spreadsheet. No coding required.

Search GitHub for "tabulapdf tabula" or visit tabula.technology. It has about 7,500 stars.

One honest warning: the last official release came out in 2018. It still works, and journalists still rely on it every day, but it is no longer actively updated. I would rather say that now than have you discover it later.

2. OpenRefine: clean up a messy list

Every newsroom has that spreadsheet from you-know-where. The subscriber list where the same street is spelled six different ways. The advertiser file with duplicates. The submitted data where half the entries end with an extra space.

OpenRefine is a free desktop program built for exactly that job. It identifies entries that are probably the same thing and lets you combine them with a single click. Data reporters often reach for it first, and it requires no coding.

It has about 12,000 stars, and it was updated within the last few days, so it is very much alive.

3. Klaxon: get told the moment a government page changes

This one was made by journalists at The Marshall Project, and it is the most newspaper-like tool on the list.

You give it web pages you want to watch. The city council agenda page. The county bid postings. The school board minutes. A state licensing or inspection database. Whenever anything on that page changes, Klaxon sends you an email.

Consider what that means on a Tuesday. Instead of repeatedly refreshing the city's website and hoping to catch the agenda when it appears, you receive an email the instant it is posted, and you are the first person making calls.

The Washington Post, the New York Times, and the Texas Tribune all use it. About 680 stars, updated this month.

The honest catch: Klaxon is not something you just double-click and run. It must be installed somewhere so it can operate continuously. This is the one where you send the link to whoever manages your website and ask whether they can set it up. It is worth asking.

Two more when you are ready

OCRmyPDF takes scanned PDFs—the kind where the text is really only an image—and places a searchable text layer underneath. Point it at your scanned back issues, and decades of your archive suddenly become searchable. Nearly 35,000 stars and updated last week. It runs from a command line, so think of it as a "get help once, then use it forever" tool.

Whisper standalone for Windows transcribes audio, and this edition comes packaged as a plain .exe so you do not need to install anything technical. Feed it a recording of the council meeting, and it returns a transcript.

Three cautions before you download anything

It is free code from strangers. Most of it is excellent. Some of it is abandoned, and a small amount is junk. The three checks above—stars, last updated, and a real README—eliminate almost all of the bad stuff.

Start with the best-known tools. A program with 10,000 stars has been examined by many people. A program with four stars has not.

Ask before installing on a work computer. If your paper has anyone who handles IT, let them know first. It is polite, and it helps you avoid a bad afternoon.

Why this is worth a Saturday morning

The tools that solve newspaper problems are usually not sold in a box with a salesperson attached. They were built by people at other newsrooms who had the same problem you have, solved it, and shared it.

No one is going to advertise them to you. That is exactly why most publishers never discover them.

Open GitHub, search for one thing that frustrates you every single week, and sort by stars. In the worst case, you spend twenty minutes and learn what a repository is. In the best case, you find something that gives you back an afternoon.