About EP Toolkit
Hey! I'm an independent developer running 
Randomware Audio and a Teenage Engineering enthusiast. What began as a browser-based tool for exporting EP projects to DAWs has grown into a full-featured desktop application.
FAQ
Is there a free version?
Yes! The browser-based version is free and open source.
Does it work offline?
Yes, the application needs internet only for license activation and updates. After that, you can use it completely offline - perfect for studios without internet access or on-the-go.
How is the license delivered?
Payments and licensing are handled by Lemon Squeezy . After purchase, you'll receive a license key via email. Enter it in the app to activate - no account creation needed.
I want [this] feature, when will it be available?
I'm always looking to improve the app and add new features! You can share your ideas and feedback through the contact form or join the Discord community to discuss features and stay up to date. I read every message and take all suggestions into consideration for future updates.
How do I get a refund?
Full refund within 10 days of purchase, no questions asked. Just send an email to ep133todaw@proton.me or use the .
How does silence trimming work?
The app uses a window-based peak detection algorithm to find and remove leading silence:
- The audio is divided into 5 ms windows.
- The peak amplitude is computed for each window.
- The app finds the first window whose peak exceeds -50 dBFS (about 0.3% of full scale).
- Everything before that window is trimmed.
This catches very quiet lead-ins that are effectively inaudible but still cause a noticeable delay when you hit a pad. Unlike simpler amplitude-only methods, the windowing approach is robust against brief noise spikes and consistently trims only the silent portion.
How does normalization work?
Normalization is a peak normalization to 0 dBFS (full scale):
- The app scans the entire waveform and finds the maximum absolute sample value.
- If the peak is below full scale (i.e. not already at 0 dBFS), it calculates a gain factor:
gain = 1.0 / peak. - Every sample is multiplied by this gain and clamped to the range [-1.0, 1.0].
This makes the sample as loud as possible without introducing any clipping or distortion. If the sample is already at full scale, it is left untouched.