Build a GitHub Portfolio: The Easiest Way to Convince Clients

The message came in at 11:40 on a Tuesday night: "Do you have any past client work I can look at?"

I didn't. Not one paid project. What I typed back was, "I don't have client work yet, but here's my GitHub, take a look at the code itself." Nineteen minutes later: "This looks solid actually. Let's talk rate." That project paid Rs 7,600. It was the first time anyone had paid me to write code, and it happened because of four repositories that no client had ever asked for.

---

The actual problem: no proof, no clients

I wanted to start freelancing as a web developer. Every platform, every cold pitch, ran into the same wall: "can you show me something you've built for a real client?" And I had nothing, because getting that first client required exactly the thing I didn't have.

I had college assignments. A couple of half-finished hackathon entries. None of it looked like something a business owner would trust with their actual website.

So I stopped waiting for permission to build something worth showing.

---

Four projects, three that actually held up

Over three weekends I built four small applications, each solving something specific rather than following a generic tutorial.

One was an expense tracker built in React, with actual data persistence, not just a fake demo state. Another was a booking system for a fictional dog grooming service, complete with time slot selection and a confirmation screen. A third was an invoice generator that exported to PDF properly, which took longer to get right than I expected.

The fourth one, a recipe search app, never really worked past the halfway point. I eventually removed it from my profile entirely. Showing something broken does more damage than showing nothing at all.

---

The README is what actually gets read

This is where a lot of people undersell good code. Clean code with no explanation is invisible to someone who isn't a developer themselves.

My format settled into something simple: two lines on what the project does, a tech stack list, one clear screenshot, and a live demo link wherever I could deploy one for free. The booking system went up on Vercel within an evening.

Actually, thinking back, that live link mattered more than the code did, for that first client at least. He told me afterward the demo link is what he clicked first. He opened the repository itself almost as an afterthought.

---

How the first client actually happened

I posted something short on LinkedIn: "Started freelancing as a web developer, here's a few things I've built [GitHub link]." Nothing elaborate.

Two days passed with no response. Then, oddly, a small business owner running a pet grooming service (the exact niche I'd invented for my practice project, which felt a little strange) sent a message asking if I could build something similar for her real business.

We talked rate. Settled at Rs 7,600 for a scaled-down version with the essentials. I quoted a five-day delivery and it took eight, mostly because the payment gateway integration fought back harder than expected.

---

What I got wrong early on

For the first two weeks, none of my repositories had a README at all. Just raw code, pushed and forgotten. Someone browsing my profile (not even a client, just someone exploring) left a comment that said the code looked fine but he had no idea what it actually did. Fair criticism, honestly.

The second mistake was worse: commit messages like "fix," "update," and once, at close to 2am, just "asdf." If a client scrolls through your commit history and sees that, it reads as careless even if the underlying code is fine.

Now every commit gets something specific, like "Add validation for empty booking slots." Small change, different impression entirely.

---

Where things stand now

It's been about seven months. I keep four repositories active, each with a complete README and at least one screenshot. I've applied the same approach to smaller, unrelated skills too, treating content writing samples the same way: real work, clearly presented, not buried behind vague descriptions.

Total earned from web development freelancing through this GitHub-first approach: roughly Rs 58,400 across several small clients. Not a huge figure yet, but it's the number that only became possible once there was something concrete to point people toward.

To be fair, a polished GitHub profile alone doesn't summon clients out of nowhere. You still have to post, still have to reach out, occasionally still have to send a cold message that might get ignored. GitHub only does its job once someone is already curious enough to click through and check if you're actually good.

---

Started a new project last night, a small library management tool for a friend still in college. Doing it for free for now. It'll go up on GitHub with a full README once it's done, same as everything else. No idea yet whether the next paying client comes from that one or from somewhere I haven't thought of.