AnySound: Make, Save and Download AI Sound Effects
A practical guide to AnySound: writing a sound prompt, loop, BPM and key settings, credits, private results and interrupted tasks.
AnySound turns a short text description into a sound effect. Describe one sound, set loop, BPM and key if you need them, then generate, listen and download the MP3. It makes one sound or a short loop at a time, not a full song.
Make your first sound
- Open the sound effect maker. Write a prompt, or press the wand next to an example to load its prompt and settings.
- Turn on Seamless loop for ambience or backgrounds. Set Tempo (BPM) and Key when the sound has to fit a track; otherwise leave them on Auto and Any.
- Sign in to generate. Your account must have a verified email.
- Follow the task status. When it is ready, play each take next to the form and select Download to save the MP3.
Loading an example or a preset never submits a generation. The model aims for the tempo and key you set but does not always hit them exactly, so check a loop before you line it up with a project.
What the current studio supports
| Control | Available values |
|---|---|
| Prompt | Up to 500 characters |
| Loop | On or off |
| Tempo | Auto, or 1–300 BPM |
| Key | Any, or a major or minor key |
| Output | MP3 files, usually more than one take per task |
There is no length setting and no WAV export. One-shot effects usually last a few seconds; ambience and beat loops often run 10 to 20 seconds.
Credits and refunds
Each sound costs 3 credits. New accounts get 20 credits on their first sign-in, enough for six sounds. If a generation fails, its credits are returned automatically and the refund appears in your credit history.
Saved results, privacy and interrupted tasks
Your prompt and settings are sent to our AI generation provider. Task records and finished MP3 files are saved privately to your account. The examples on the homepage were created by the AnySound team, not taken from user generations. Read our privacy policy before submitting sensitive material.
The results list shows your latest 20 tasks. Play or download a finished sound again at any time from that list. Only one sound can generate at a time per account.
If a connection drops, reload the page; an unfinished task keeps updating on its own. If a submission could not be confirmed, it is marked as unconfirmed and is not charged twice. Contact support@anysound.org before resubmitting, because the provider may already have accepted it.
Write a focused prompt
Name the sound source, the action, the space and how the sound ends, then say what to leave out. For example: “A heavy wooden door slams shut, low thud with a short natural decay. Close-miked, isolated sound effect, no music.”