What's New

Sound engineers know this particular kind of search: you need a door creak — not just any creak, but that one, heavy, cellar-deep — and you are on the fortieth file, no longer sure how it differs from the first. Now a sound is not something you look for. It is something you describe.

We built a tool that makes sounds out of descriptions. You type “a hedgehog snuffling”, “a bartender wiping a glass”, “a car alarm at four in the morning” — in English, in your own language — and you get a sound. Not speech: it cannot pronounce words, its nature is a different one, it is about rustle and impact, about rain and creak, about what lies between words. It runs on tokens. You can download it as MP3, or as lossless WAV — 48 kilohertz, the way an editing suite expects it.

Use it in your own work: video, games, apps, podcasts, advertising.

Short effects run up to twenty-two seconds and can loop without a click — for backgrounds and game loops; long atmospheres run up to three minutes. Whatever comes out stays in your library, so you are not digging through folders later.

About the terms — plainly, no fine print. You may use the sounds in your own work, including commercially and with no time limit. You may not sell or distribute the sounds on their own — as packs, libraries or stock collections; only inside a finished work. There is no exclusivity: the same prompt from someone else will give a similar sound.

Seven examples with players are waiting on the tool page. A rooster, a thunderstorm, crying, a hedgehog, the bartender with his glass, a car alarm, a cat. It is easier to listen once: Sound Effects Generator.

Tools that studios used to pay for should today be within everyone's reach. A person with a phone and an idea shouldn't have to yield to a person with a budget. That is what we work for: hundreds of tools, and a new one almost every day.

But useful things have no loud voice. Places like ours are heard of last — the one heard first is not the one who is needed most, but the one who paid most to be heard.

You have what we cannot buy: a voice people trust. One sentence from you — «this is what I actually use» — weighs more than a month of advertising. Help us be heard, and we will thank you with what we have.

You — tell your people about us. In your own words. Honestly.
We — give you a Premium account for six months: higher daily limits across all tools, server-side processing of long recordings, priority on heavy jobs, exports at maximum quality, no ads and no watermark on video. Plus everything new that ships during those six months.

Who this is for. Our benchmark is 10,000 followers or a comparable audience: viewers, listeners, community members. If yours is noticeably smaller we will most likely say no — and we would rather not waste your time. One exception: a narrow professional audience — sound engineers, teachers, 3D-printing makers. There we look at the topic, not the counter.

We will not tell you what to say. We will not send you ad copy and we will not ask you to praise us. Say it in your own words and honestly: if you did not like a tool or something is missing, say so. That is more useful to us than praise.

Where to write. moc.acirbmet@troppus, with «Blogger» in the subject line. Attach links to your platforms and a couple of words about which of our tools you use. We read every message ourselves, so a reply may take a few days.

And if you do not have a large audience — just show the site to one person who would find it useful. Everything starts with one person.

— The Tembrica team

A new tool is now open — Text Behind Image. A neural network separates the person or object from the background and the lettering slips behind it, giving the sense of depth you see on posters and album covers. It works on photos and on video.

All five are genuine output from the tool: a neural network cuts the subject out and the lettering runs behind it.

We also reworked the editor: the sliders are easier to see and to hit, text layers can now be reordered with the arrows or by dragging, and the settings panel adapts to the space available — a portrait shot no longer sits between empty side margins.

Now a question for you. Lately we have spent a lot of time on photo tools. Should we keep going that way, or turn to something else?

If you picked "Something else", tell us in the comments what exactly. And if you have a specific task you currently solve by hand or in some heavy program, describe it. Messages like that are what most of our new tools grow out of.

— The Tembrica team

We build a lot of tools and try to release something new almost every day. We don't post about all of them here — so this page stays readable and the things that really matter don't get lost in the stream.

A tip: use bookmarks. Every tool has an "Add to bookmarks" button — save the ones you need and they'll always be at hand, so you won't have to dig through our 560-plus tools each time.

And there's more coming. By the end of the year we want to reach a thousand tools. All of it is included in the Premium subscription — you won't pay extra for each new tool.

Now, honestly — we can't release everything. Two powerful tools are finished and fully working, yet we won't publish them.

The first is voice cloning: you make a digital copy of your own voice, type some text, and it speaks in your voice. We meant it for people who narrate audiobooks. But the deeper we looked, the clearer it became: the same voice can be used to impersonate someone else. We did a lot to guard against abuse, and some of it worked — but the technology is still powerful enough to cause harm. That's a responsibility we're not taking on right now.

The second is photo animation: to bring an old photo of your grandmother or grandfather back to life. A warm idea — but here too it turned out that such technology is far too easy to turn against ordinary people and into a tool for those with bad intentions.

Here's something new that will harm no one, though — Synonyms. It'll keep growing, and we hope it's useful to you.

— The Tembrica team