Everyone’s Using AI to Write Resumes—Here’s a Tool That Lets You Speak Yours

admin4 hours ago

0 9 7 minutes read

I build a new free tool for Resumémaker.onlineAnd I think it is more robust and has more potential than all other free tools that I have manufactured so far.

This is why I thought it would be a good idea to share the process as you go.

The idea is simple: I want to help people create short 30 -second video CVs.

Not to replace their CVS, but to stand out a little.

Here is the problem I see today: everyone uses AI and has perfectly written CVs, which is great … but that means Everyone mixes.

So what happens if you could just add a short video? This could make your application much more memorable. You can reach it to your CV or publish it on LinkedIn.

Now, obviously, not everyone is comfortable with this:

People are not video editors.
Most webcam recordings seem terrible.
And many of us are shy or not excellent to talk to the camera.

This is why I have to make the easiest than possible for someone to hit a recording.

I know that there are also concerns about biases, as in some countries, you are not even supposed to add a photo to your CV … so the idea of a video may seem a terrible idea.

But again, This is not a replacement for a traditional CV. This is just a complement. A way to show a little personality, if you wish.

Anyway … a challenge at 5 days. This is the plan.

📅 Day 1: The first steps

Today, I started with the basics. I reused a code that I already had to let users download their CV, a PDF or a Word Doc, extract the text and transmit it to Chatgpt.

You can also stick a job to make the script more personalized and also choose the language.

The AI then generates a short 30 -second video script based on your CV.

The prompt is designed to appear relaxed and real, as you speak to someone. And of course, you can change the script as you wish.

I also added the webcam overview, the option to choose the video aspect report (wide, square or vertical screen), and you can already press the recording and download the video.

It's a small first step, but it works.

Tomorrow, I will attach myself to a key element in the recording process: the telepromitator.

Day 1, fact ✅
Let's see how far I can follow him.

📅 Day 2: the telepromoteur

I knew that just generating a script and letting users register on the webcam was not enough.

I wanted to build a tool that really does it Comfortable and extremely easy for people to register.

First of all, I added a simple Teleapompter with automatic scrolling text. It was the obvious decision.

But then I thought … What if I could really recognize the words you say while you are talking about?

The telepromoteur could scroll your pace and highlight Exactly where you are.

I started by checking word by word, but it was a bit clumsy and unreliable.

Tiny differences in pronunciation would discourage it, especially if you do not speak in your mother tongue (like me).

If a functionality like this is only half-work, it is worse than not having it at all. It must be reliableOtherwise, this adds an additional level of frustration: sometimes when you spoil as part of the normal process, but even worse when you were doing everything correctly and the Teleapper has delayed, forcing you to start again.

So I deleted the word detection by word and passed to the detection of sentence, which made all the difference.

There is a little magic that happens under the hood, and now it seems much more stable and reliable.

The icing on top is that The recording stops shortly after it detects the last sentence. So you don't get this clumsy break where you reach the mouse to click on stop. It ends alone, pleasant and smooth.

I also made it possible to click anywhere or press any button also stop recording, just in case you want a little more control.

For example, you can hold the mouse and simply click to stop it without breaking your pose, because a click will put an end to the recording.

A limitation: the voice recognition part only works on Android and the desktop (Windows and Mac). It doesn't work on iOS, or at least I still can't make it work.

If you are on the iPhone or iPad, it will be by default automatic scrolling, which is always available for everyone.

Everything sounds well, right? But the truth is that none of this matters if the telepromoteur always gives you the impression of reading the screen.

You are supposed to show your human side (or at least the false elegantly) do not read like a robot.

So here is what I did: I placed the televimmer at the top of the screen, centered just above the webcam.

In this way, when you are R ** EADING, it seems in fact that you establish visual contact (** especially if you get a little bit of the device).

I also kept the narrow text so as not to scan from one side to the other, which would make you obvious that you read.

And of course, You are not supposed to look at the words all the time.

I would probably add a little tip in the user interface to remind people to break the eyes from time to time and keep it human and natural.

Day 2, done! ✅
Much more complicated than I thought, but enormous progress.

📅 Day 3: Personalization options

Today, it was another key aspect: allowing users to personalize appearance and sound without the need for publishing software.

I started by letting users add their name Androle or e-mail to a Video superposition model which is on the video and makes it instantly more polite.

You can preview exactly what it will look like before hitting the recording:

I plan to Design at least 10 of them, each with a different atmosphere (from clean and minimal to daring and colorful).

Some of these models will be locked unless you download a CV made with Resumémaker.online.

It is intentional to keep any connected and to bring more users to the main paid product. It is always extremely generous given the quantity you get for free.

I also added background music Options. Nothing above, just simple tracks to make your video more polite.

You can preview them, choose one and it is integrated directly into your final recording. I would probably add at least 10 tracks here too.

Yeah, I know it always looks like shit, but that was the point where I In fact Start seeing the potential.

Day 3, made ✅
It is finally starting to feel like a real powerful tool!

📅 Day 4: Make it look good

Today was one thing: to make things good well.

So far, everything was a raw prototype, which is important when you try to move quickly and test ideas.

The first days of a project generally come with a lot of excitement and energy, But this momentum fades quickly.

This is why I try to use these first days to focus entirely on the functioning of the basic product.

For some people (and sometimes for me too), it can be discouraging to work on a product that always looks so rough. But the truth is that thereYou can waste a lot of time doing something pretty before it is really useful.

Personally, I try to get to a point where the main features are solid.

Once I'm there, it becomes much easier to stick to this and finish. I start to see that the product is useful and I get a little boost of excitement.

The difficult part is over. Now it's just a question of varnish. From this moment, he should want to go downhill.

First of all, I designed the video superposition models. They are clean, subtle and operate with all kinds of styles and in all video aspect ratios.

I am Really Happy with the way they turned out.

I have also added to police couples options so that your name and your role correspond to the visual atmosphere of the curriculum vitae model if your CV has been made to Resumémaker.online::

I designed the user interface to keep it minimal, but with some good details, such as the blurred video background.

I added useful advice at the bottom of the screen and improved the telepromote by highlighting the sentence in which you are currently, while disappearing the next.

With the new user interface, the televimmer is not centered at the beginning. But when you press the recording, the settings are hidden automatically. In this way, it seems that you look at the camera while reading (as I have already explained).

I also added 10 subtle background tracks to choose from and I made sure that everything works well on mobile with touch interactions.

It was the longest day of work in terms of hours, but also the easiest in terms of flow.

Make a solid base All Otherwise, is set up.

Day 4, made ✅
Tomorrow… the launch time!

📅 Day 5: logo, destination page and launch

Last day. Let's go back!

I designed a simple logo, a variation of the one I use Resumémaker.onlineTo keep the visual connection while giving this tool its own identity.

I also recycled parts of the layout of the home page and adapted the design of this new product. The idea is to let it live as an autonomous tool.

I created visuals and built versions in several languages. I also added the automatic-translastic support directly in the tool.

Wrote some quick blog articles related to curriculum vitae video, but other than that, it's only a page, a clean copy, a strong CTA, no stuffed animals.

So here it is. The tool is ready.

The final flow is super simple:

Download your CV
(Optional) Add a job post
Get your script
Select your model
Record
Read with the telepromoteur
Download your video CV

No registration. No connections. No assembly required. Everything in the browser.

I started to know that it would be a little more robust than my usual free tools … and now We really have the impression that it could be a product to him.

If you are looking for hunting, or if you just want to try something different to stand out, it's live: 👉 Renumaker.video

Also, the The launch of product hunting is live today (May 2). I would like your support: 👉 https://www.producthunt.com/posts/resumemaker-video

** Day 5, made ✅
**

Now let's see if people use it. I would love to hear your comments!