Skip to main content

For YouTube creators

Thumbnails that get the click, built layer by layer.

A layer-by-layer thumbnail editor. Background, subject, elements and text stay separate, so you redo one part without losing the rest.

Open the editorOr the free image tools
  • No account, no sign-in
  • No limit, no watermark
  • No image reaches a Drewvy server

The editor works with nothing but your own photo. The AI panels inside it use your own Google Gemini key — free to create, and Google meters what you use on its own pricing. Why your own key?

A screenshot of the Drewvy editor: a layer list on the left with Text, Elements, Subject and Background, a 1280 by 720 canvas in the middle showing a cut-out person on a neon background under the headline MY NEW SETUP, and the subject's position and scale controls on the right.
The editor, mid-build. Four layers, a fixed canvas, and every control for the one that is selected.

Made in Drewvy

Three exports out of the editor above, at the canvas's real 1280 × 720 — not mockups.

  • Thumbnail: a smiling man in a green shirt cut out over a blue neon light background, with the headline MY NEW SETUP in white block letters.
    Background, cut-out subject, one text layer.
  • Thumbnail: a laughing woman cut out over a concert stage lit by coloured lasers, with the headline FESTIVAL VLOG in orange block letters.
    Same three layers, a different photo and preset.
  • Thumbnail: a man with folded arms cut out over a close-up of a computer motherboard, with the headline PC BUILD 2026 in yellow block letters.
    Uploaded background, no generation involved.

The faces are stock photography, not customers: there are no users to quote yet and nothing here pretends otherwise. Each one was built the same way — a photo uploaded to the Subject layer and cut out in the browser, a second photo on the Background layer, one text layer, then Compose and Download. No AI call and no API key was used for any of the three.

How one thumbnail gets built

Four layers, generated one at a time and composed on the canvas you pick — 1280 × 720 by default — in your browser. Redo any one of them without touching the other three.

  1. Stage one: the background photo alone on the 1280 by 720 canvas — blue neon light trails over a city.
    Background. Uploaded here, or described and generated.
  2. Stage two: a man in a green shirt, cut out, composed on the right of the same neon background.
    Subject. The photo, cut out in the browser and placed on top.
  3. Stage three: the same image with the headline MY NEW SETUP in white block letters at the lower left.
    Text. Drawn onto the canvas at full resolution, then composed and downloaded.

Three stages, four layers: this build used no Elements layer, so there is no fourth picture for it.

  1. Background

    Describe one, restyle an upload, or drop in an image you already have.

  2. Subject

    Your photo, cut out in the browser. Change the pose or the outfit; your own photo is the identity reference every time.

  3. Elements

    Arrows, badges, props — generated or uploaded, each one its own object you can move.

  4. Text

    Drawn straight onto the canvas at full resolution, so it stays sharp at 1280 × 720.

Then compose, and download JPG, PNG or WebP at the canvas's real pixel size. Every thumbnail on this page is a real export out of this editor, and the screenshot at the top is the editor itself. There are no testimonials here, because there are none to show yet.

Five AI tools, free, driven from your browser on your own key

These five are free and need no account. The one thing you bring is your own Google Gemini key: free to create, kept in your browser, and metered by Google against your own Google account. The seven tools that need nothing at all are the next section.

Editor

Keep the parts that work, redo the rest.

Background, subject, elements and text are four separate layers, each one moved, scaled and regenerated on its own, on any canvas from YouTube HD to 4K, square or vertical, exported as JPG, PNG or WebP.

Inspire

Copy the look, keep your own face.

Paste a YouTube link or upload a thumbnail you like, and Drewvy rebuilds that scene at 16:9 with your own face cropped from your own photo.

Thumbnail Variants

Give one video three covers.

Pick an image and get up to three rearranged versions of it — alternates to try, not click predictions, because the testing is still yours to do on YouTube.

Background Swap

New background, same you.

The default mode cuts you out on your own machine and only replaces what is behind you; the AI mode repaints the whole frame, objects and text included, and says so before you use it.

Profile Picture

One selfie, every platform size.

Styles from professional to anime, previewed round or square, downloaded at the exact size each platform wants.

Also in the editor: clothes fusion from an outfit photo, pose changes and eight one-click background styles — AI calls that go from your browser to Google on your key. Saved subject presets, frames, the composition grid and saved projects are browser storage, and stay in it.

Not built yet, and labelled that way: presentation graphics, web content assets, brand kits, batch generation.

These five need your own Google Gemini key — free to create, and Google meters what you use. Why it is free

Seven image tools that need nothing at all

These need no sign-in and no API key — nothing but the image. Six of them run entirely on your machine; the seventh just fetches a YouTube thumbnail from its link. No watermark, and no cap on how many images you run.

Six of these never send anything anywhere. The seventh sends a YouTube video id to fetch that thumbnail, because YouTube blocks the browser from doing it. Background removal downloads its model the first time you remove a background — about 100 MB, or 111 MB if your browser runs it on the GPU — and then keeps it cached.

Where your images actually go

Drewvy has no image server. There is no upload endpoint, no account and no database at all. That is not a policy, it is how it is built.

What stays on your machine (4 things)

  • Your photos, cutouts, subject presets, projects and finished thumbnails live in your browser's own storage. Open your browser's developer tools and you can see them under these names; the subject store still carries the name it shipped with. drewvy-gallerydrewvy-projectsdrewvy-characters

  • Your Gemini key lives in your browser and is sent only to Google. localStorage

  • Background removal, green screen, cropping, resizing, compressing, text and the final composition all run on your device.

  • In Background Swap's default mode the cutout of you is composed locally, so your own pixels are never repainted and never uploaded.

Everything that leaves, said plainly

  • When you use an AI tool, your browser sends that image straight to Google's Gemini API with your key. It goes to Google, not to Drewvy.

  • When you paste a YouTube link, Drewvy's server fetches that thumbnail, because YouTube blocks the browser. Only the video id crosses the network — none of your images are involved.

  • If you choose to leave a tip, that happens on GitHub Sponsors, so it goes through your GitHub account and GitHub receives your payment details. Drewvy never sees your card, and nothing about the site changes either way.

  • Every page loads an Umami analytics script, which counts page views.

That is the complete list. If a claim is not on it, I am not making it.

Your browserYour photoThe cutoutProjectsYour API keyGoogle Geminithe image + your keyDrewvy servera YouTube video id, no image
The only two paths that can carry an image, or a request about one, out of your browser. Everything else in the list above is a tip payment or a page count, never picture data.

Drewvy is one person and no image server

Drewvy is built by one person.

It can be free because it barely costs anything to run. Your own machine does the image work. The AI calls go from your browser straight to Google on your key. There is no image server here. Nothing of yours to store, and nothing for me to pay per image.

So there is nothing to buy. If it saved you an evening and you want to say thanks, you can — and if you never do, nothing about the site changes.

What that means for you

  • Every tool is free. There is no licence, no key to buy, no trial and no limit on how much you make.
  • No account. Nothing on this site asks you to sign in.
  • The five AI tools run on your own Google Gemini key, so your generations are yours and Drewvy never bills for an image.
  • A tip buys nothing. Nothing here is held back for anyone who leaves one, and no tool behaves differently afterwards.
Leave a tip, if you want to (opens in a new tab)

Opens GitHub Sponsors in a new tab. Nothing here is locked, so a tip changes nothing about the tools — it is a thank-you, not a purchase.

The other way costs nothing: tell one person who makes videos that this exists.

Two steps to use it, and no third one

There is no account step and no payment step. The only thing you bring is a Gemini key, and only for the five AI tools.

  1. Open any tool.

    No sign-in, nothing to install. The seven Free Image Tools work straight away, on your own machine.

  2. For the AI tools, paste your own Gemini key.

    One Google page, a few minutes, no card. The key stays in your browser and is sent only to Google, which meters what you use on its own pricing.

    Get a key at aistudio.google.com/apikey (opens in a new tab)

  3. Make as much as you want.

    No cap, no watermark, no per-image charge from Drewvy. If it ends up earning it, there is a tip link above — it unlocks nothing, because nothing is locked.

I would rather ask you once than bill you every month. Nothing here bills you at all, whether or not you ever leave a tip. I do not intend to put any of this behind a payment.

Questions before you start

Is my face uploaded anywhere?

Not to Drewvy. There is no image server and no image in a database. The free tools work entirely on your machine. When you use an AI tool, your browser sends the image straight to Google with your key, so it reaches Google and nobody else. In Background Swap's default mode, even Google never sees the cutout of you.

Is it really free?

Yes. Every tool, with no account, no licence, no trial, no watermark and no cap on how many images you make. The only thing you bring is your own Google Gemini key for the five AI tools, and Google bills that to you, not through me.

Do I need an account?

No. There is no sign-up, no sign-in and nothing to activate. Open a tool and use it.

Why do I need my own Google key at all?

Because the alternative is me running a server that holds your photos and reselling you credits with a markup on top. Your key removes both. There are no credits to buy, no markup, and your usage stays yours.

Get a key at aistudio.google.com/apikey (opens in a new tab)

Do I have to pay Google?

Creating an AI Studio key is free and takes no card. After that Google meters what you use against your own account under Google's current pricing, which is Google's to set and change, not mine. I never bill for generations and I do not sell credits.

What if I do not like what it makes?

Generate again. Drewvy puts no limit on regenerations and never charges per image. In the Editor you redo one layer at a time, so a bad background never costs you the face and the text you already liked. Expect Google's safety filter to refuse a prompt now and then, usually when a brand or a real person's name is in it — rephrasing usually clears it.

Does it work on my phone?

The free tools are built to work on a phone. Profile Picture mostly does too. The Editor wants a big screen: four layers, a canvas and a properties panel do not enjoy 360 pixels. Background removal downloads its model on first use, so start that on wifi.

What does a tip get me?

Nothing extra, and that is the point. There is no premium tier, no unlock and no feature waiting behind a payment. A tip is for the building, if the thing turned out to be worth something to you.

Tip on GitHub Sponsors (opens in a new tab)

How long does one generation take?

Each one is a single call from your browser to Google. Drewvy waits up to two minutes and then lets you retry, and it retries rate limits and server errors for you.

Support DrewvyOpen the free tools

Use all of it for as long as you like. Tip only if it has earned it.