Vesper was the name I gave to a browser I tried to build. I had no coding knowledge, I did not know how a browser is made, and I had never downloaded a project of that size. I had a laptop, a lot of curiosity and AI chats to guide me. The project ran for about 45 days and I never released it. This is the whole story, told as honestly as I can remember it.

The idea: a browser you talk to

At the time I was experimenting with AI and with running models locally. I kept trying small things, and one idea stuck: a browser that works completely by voice. No keyboard, no touching the laptop. The AI models would run on the laptop itself, so the cost would be zero and there would be no worry about losing data to someone else's servers.

I imagined saying something like "go to YouTube, search for this, scroll, click that video, play it and skip the ad." The AI would then use the mouse and keyboard instead of me. I also planned web search with summaries, shopping, booking and, in my own words at the time, every kind of task.

It was meant to be for my own use first. Whether it would ever be public was something I planned to decide later.

Starting from zero

I will be direct about where I began. I did not know how to code. I did not know how a browser is built or how to download Chromium. At that point AI chats were the only guide I had. My machine was a Lenovo laptop with an Intel Core i7 processor, and I do not remember the exact specifications.

The first month went into discussions and small experiments, trying different ways to build a browser. Then I learned that Chromium, Google's open-source project, is the root of Chrome and of many other browsers. That changed the plan: I would not write a browser from nothing, I would start from Chromium and put my own interface and branding on top of it.

I tried Firefox-based options as well. Chromium felt hard at the start and easier later, while Firefox felt easy at the start and harder in the future. Most people who build browsers also start from Chromium, so I chose the same root.

Why the name Vesper

The name came from Comet, the browser from Perplexity. I liked how clean and unique that name was and wanted something with the same feel. While searching I found Vesper, which is an old word for evening. I liked the picture it gave me: sitting with your laptop in the evening and getting your work done without typing a single command.

Before I downloaded anything, I had already spent a few days deciding the name and planning the logo.

The download that tested my patience

I first read that Chromium would take 30 to 70 GB. It turned out to be much more, over 100 GB. I planned one month of unlimited mobile data for it and started the download.

The first two to four days failed. The hotspot disconnected, downloads stopped and data was lost. Sometimes the laptop shut down because the charge hit zero. Then I restarted everything. The laptop stayed on around the clock, and across the project it was running 24 hours a day for about a week in total. This cycle repeated for roughly 15 to 20 days before Chromium was fully installed on my laptop.

I used Ubuntu for cloning and some of the work, along with VS Code, the terminal and AI chats. I did not have a hardware problem. The laptop stayed smooth, although it sometimes slowed down.

The day it actually built

The build completed, and I had a working app. I had changed chromium.exe into vesper.exe with my own icon and logo, and when I opened it, my own interface appeared. I changed the name, the icons, the logos and much of the front end. Underneath, the core of the browser is written in C++ and other code that I never learned.

It is important to say what I did not build. I only changed the front end: the design, the look and the branding. The voice control, the local AI, the search, the shopping and the booking were never built. No privacy feature was built either.

AI did most of the technical work. I would estimate around 60 to 70 percent. My part was execution: running commands, copying errors back to the AI and asking for the next step. I had zero knowledge and a hundred percent curiosity.

The wall: errors, and then the backend

Every change I made produced an error. Out of every ten commands the AI gave me, only about one solved the problem. That was the most frustrating part of the whole project, and it happened again and again.

Then I started working on the backend and realised how big the project really was. Browsers like this are built by teams of many developers over months or years. I was alone, working with an AI that gave the right fix about once in ten tries. A lot of time was wasted, and my motivation to continue slowly faded. I also kept asking myself how I would maintain something this large by myself.

No single error ended the project. I stopped because the work got slower, a new idea arrived, and eventually I removed it from the laptop.

What I took from it

The best part was the curiosity itself: stepping into one of the biggest open-source projects in the world, poking around and changing things. Oddly, it became my best building phase only after the project failed and I dropped it. It was also the first time I kept a laptop running 24 hours a day for about a week, and the first time I used more than 300 GB of data.

The biggest technical lesson was how to prompt AI to solve errors. By my second month I noticed roughly a 10 percent or better improvement from the same model, only because I had learned how to ask. Before Vesper I used a few AI tools and templates to make websites. After it, I started using AI in every task, decision and piece of work.

It also showed me what the start of an entrepreneur's journey can look like: running experiment after experiment, and a kind of curiosity that keeps growing.

Where Vesper stands today

Vesper was an experiment, and I stopped it during the building phase. It was never released and I do not have any files left, so there is nothing to download. The name and this story are what remain. A short project summary also sits on the Work page, together with my other projects.

Questions about Vesper

What was Vesper?

Vesper was the name of an AI browser that I planned and started building. The idea was a browser controlled completely by voice, with AI models running on my own laptop.

Can I download Vesper?

No. It was never released and I removed it from my laptop after about 45 days, so there is no download.

Was Vesper built from scratch?

No. It was built on Chromium, the open-source project behind Chrome. I did not write a browser engine. I changed the name, the icons, the logo and the interface.

Did the voice and AI features ever work?

No. I only worked on the front end, meaning the design and branding. The voice control, local AI, search, shopping and booking were planned but never built.

How long did it take and how much data did it use?

The project ran for about 45 days and used more than 300 GB of data. About 15 to 20 of those days went into getting Chromium downloaded and installed, mostly because of failed downloads.

How much of the work did AI do?

Most of the technical work. I estimate around 60 to 70 percent. My part was execution, running commands and reporting errors back, with no coding knowledge at the start.

Why did you stop?

When I reached the backend I saw how large the project was for one person working with an AI that fixed errors about once in ten tries. Time was wasted, my motivation faded and I decided to stop.

Will you build a browser again?

No. My direction has moved towards entrepreneurship. If I ever did something this large again, I would work with a team and manage my time better.

What does the name Vesper mean?

Vesper is an old word for evening. I chose it because I liked the idea of sitting with a laptop in the evening and getting work done without typing commands.