3 Free Tools to Add Pro Sound Effects to AI Video Edits
The short answer: use three free tools with your AI agent. Openverse API finds real, openly licensed sound effects. Real SFX places sounds on the exact frame of your edit. Earshot makes a new sound in code when nothing else fits. Together, a silent AI edit ends up sounding like a person with a good sound library cut it.
AI can cut a video in minutes, but the sound is usually where it gives itself away: silent transitions, generic whooshes, hits that land a few frames late. These three tools fix exactly that. Below is what each one does, when to reach for it, and how to set it up.
The 3 tools at a glance
- Openverse API: finds real recorded sound effects that are free to use.
- Real SFX: reads your edit and puts real sounds on the exact frame.
- Earshot: makes a brand new sound in code when you can't find one.
In plain words: an API is a door that lets software talk to a website directly. A skill is a folder of instructions that teaches Claude Code a new job. An MCP server is a plug that lets an AI agent use an outside tool.
1. Openverse API: find free sound effects
Openverse is a free, open search engine for openly licensed images and audio, run by the WordPress community. Its audio comes from sources like Freesound, Jamendo and Wikimedia Commons. The API lets your agent search it and download sounds itself, instead of you browsing sound sites by hand.
When to use it: when you need a real recording, such as a door slam, a crowd, rain, footsteps or a camera shutter. Real recordings sound more natural than generated ones for anything physical.
How to use it:
- No sign-up needed to start. Search with a plain web address like
https://api.openverse.org/v1/audio/?q=whoosh&category=sound_effect&license=cc0. - Or just ask Claude Code:
Search the Openverse API for CC0 whoosh sound effects, download the 5 best short ones into a folder called sfx, and save the title, creator and licence of each in credits.txt. - For heavier use, register free credentials through the API's auth endpoint. Anonymous requests are capped at 20 results per page and are rate limited.
Watch the licence: filter with license=cc0 if you don't want to credit anyone. CC BY sounds are fine but need a credit line, and anything marked NonCommercial is off limits for client or paid work. Openverse's terms also ban scraping, so search for what you need rather than bulk-downloading.
Link: Openverse API → api.openverse.org/v1
2. Real SFX: put the sound on the exact frame
Real SFX is a free Claude Code skill by Clipwing. It reads your video, finds the moments that need sound (cuts, pops, text landing, clicks, typing), picks real recordings from your own sound folder and places each one on the exact frame. It keeps your music and voiceover, and hands back a new MP4 plus a separate sound effects track you can adjust in any editor.
When to use it: when your sounds exist but land off-beat, or when you have a finished render and a folder of sounds and want them placed properly without doing it frame by frame.
How to use it:
- Install Python 3.9 or newer and FFmpeg (a free tool for processing audio and video).
- Copy the
real-sfxfolder into~/.claude/skills/and restart Claude Code. You can also install the.skillpackage from the Claude app with Save skill. - Point it at your sounds:
python scripts/setup.py init --library main="D:/Sounds/SFX" --uisfx --index. The--uisfxflag adds 78 free interface sounds if you're starting from zero. - Ask Claude:
Add sound design to this render. It returns the new video, the sound track and a report of every sound it placed.
Good to know: Real SFX never generates sounds, by design. If your folder has nothing that fits a moment, it leaves that moment quiet and tells you. Claude also can't hear, so it checks timing by measurement and flags a few moments for you to listen to before you publish.
Link: Real SFX → github.com/Clipwing/skills
3. Earshot: make a new sound in code
Earshot is a free, open-source toolkit that gives coding agents ears. It can describe a sound, check whether a sound matches a description, and make new sounds in code with no recording involved. It has a built-in bank of 252 sounds (whooshes, pops, cut hits, clicks, camera shutters, glitches and more) that render by name without the internet.
When to use it: when you searched and found nothing, or you need motion graphics sounds like risers, swishes and UI pops in an exact length. It's also handy for letting your agent check that a downloaded sound really is what its title says.
How to use it:
- Install Python 3.10 or newer and FFmpeg. It runs on a normal computer, no graphics card needed.
- Clone it with
git clone https://github.com/Leonxlnx/earshotand follow the install steps in the README. The first run downloads about 1.1 GB of models. - Connect it to Claude Code:
claude mcp add earshot --scope user -- /path/to/earshot/.venv/bin/python -m earshot mcp. - Make a sound, for example a 0.8 second whoosh:
earshot sfx render whoosh_fast/crisp -o w.wav --dur 0.8. Or just ask Claude for the sound you need.
Link: Earshot → github.com/Leonxlnx/earshot
Using all three for sound effects in AI video editing
- Find: ask your agent to pull the real sounds you need from Openverse into one folder, with credits.
- Place: point Real SFX at that folder and let it put each sound on its frame.
- Fill: for any moment Real SFX left quiet, have Earshot make the sound, then drop it in at that timestamp.
- Listen: play the final cut once with headphones before you post. This step is still yours.
FAQ
Are these sound effects tools free?
Yes. All three are free. Openverse is free to search, and Real SFX and Earshot are free, open projects on GitHub. What you pay attention to is the licence of each sound you download from Openverse.
Do I need to know how to code?
Not really. You'll copy a few commands during setup, and after that you talk to Claude Code in plain English. If a command fails, paste the error into Claude and ask it to fix the setup.
Can I use these sounds in client work?
Sounds marked CC0 or Public Domain, yes, with no credit needed. CC BY sounds need a credit line. Skip anything marked NonCommercial for paid work. Sounds Earshot makes are generated fresh, and its code is under the MIT licence.
In short
Openverse finds real sounds, Real SFX lands them on the right frame, and Earshot makes whatever is missing. Set them up once and your AI edits stop sounding silent or synthetic, as long as you still give the final cut one honest listen.
Quick summary · key points
- ✓The 3 tools at a glance
- ✓1. Openverse API: find free sound effects
- ✓2. Real SFX: put the sound on the exact frame
- ✓3. Earshot: make a new sound in code
- ✓Using all three for sound effects in AI video editing
- ✓FAQ