Welcome to Tapechild — text-to-sample for your DAW
Tapechild turns plain English into drag-and-drop samples for Ableton, Logic, FL Studio, and Bitwig. Runs on-device on Apple Silicon. No cloud, no plugins.
We built Tapechild because the AI music tools we wanted didn't exist yet.
Every "AI sample generator" we tried was either a web app you had to babysit, a plugin that broke after every DAW update, or a cloud subscription that uploaded your prompts to someone else's server. None of them fit the way producers actually work — describe a sound, drop it on a track, move on.
So we made the one we wanted.
What it does
Type a sound. Get a sample. Drag it into your DAW.
That's it. No render queue, no plugin to install, no preset hunting:
- Text → sample. "Dusty 808 with tape hiss". "Warm Rhodes loop at 92 bpm in F minor". You get a clean WAV in seconds.
- Drag into your DAW. Every result is a real file on disk. Drag the card straight onto a track in Ableton, Logic, FL, Bitwig. No VST. No bridge.
- One-shots & loops. Switch between single hits and seamlessly tileable loops. BPM and key auto-detect, or pin them yourself.
What's different about it
It runs on your Mac. We built the generation engine on Apple's MLX framework, which means everything happens on-device on Apple Silicon — no cloud, no upload, no queue. Your prompts never leave the machine. The flip side: you do need an M-series Mac.
No subscription tax for the model itself. Cloud generators charge you per render because their costs scale with usage. Ours don't. You buy the app, the GPU is already in your laptop, and you generate as much as you want.
Royalty-free, commercial use. Anything you make with Tapechild is yours. Put it in your beats, your client work, your label release. No attribution required, no sample-clearance nonsense.
Who it's for
Tapechild is built for producers who want to spend less time hunting for sounds and more time arranging them. If your workflow looks like "I need a specific thing, right now, in this session" — that's the gap we're filling.
We're early. The Mac app is shipping today. Windows is on the roadmap once the MLX → ONNX path stabilizes. If you want updates, grab the early-bird license — it locks in the price and gets you every model and feature update for the year.
More posts coming: how the on-device engine works, prompt patterns that get the best results, and a deep dive on why we chose MLX over Core ML.
Talk soon.