A variety of tools and 3D models and tools I’ve created for Unity, designed to help speed up development and enhance your projects. Available now on the Unity Asset Store. Whether you’re working on a game, prototype, or interactive experience, these assets are designed with flexibility, performance, and ease of use in mind.
A growing collection of themed voice clip packs designed specifically for Unity. Each pack includes high-quality, hand-crafted voice lines tailored to different categories—such as characters, combat, narration, and more. These assets are ideal for quickly adding personality, emotion, and clarity to your games or interactive experiences. Easy to implement and organised for fast integration, they’re a great way to bring your project to life with voice.
A selection of 3D models available outside the Unity Asset Store. This collection includes a mix of assets—some optimised for real-time use in games, while others are high-poly presentation pieces suited for animation, rendering, or concept work. All assets are hosted on external platforms and available for direct download.
Stop digging through folders.
Let AI tag your entire audio library so you can find any sound in seconds.
You know the feeling.
You need a specific sound — a dry footstep, a tense low drone, a punchy one-shot hit — and you know it’s somewhere in your library. Twenty minutes later, you’re still digging through folders named “SFX Pack Vol 3” and “Misc Sounds Final Final.”
Audio Library Tagger fixes that.
Point it at your audio folder. Walk away. Come back to a fully-tagged, fully-searchable library in your browser.
How it works
Audio Library Tagger runs two AI systems on every file in your library:
PANNs (Pretrained Audio Neural Networks) — Google’s open-source model trained on 527 sound categories. Detects explosions, rain, footsteps, pianos, engines, crowds, birds, impacts, whooshes, and hundreds more — with confidence scores. Runs on your NVIDIA GPU for serious speed.
librosa feature analysis — Estimates BPM, musical key (e.g. A minor), loudness, brightness, and texture. Automatically tags each file as fast/slow, loud/quiet, bright/dark, one-shot/loop/full-track.
Everything goes into a local database. A clean browser-based search UI lets you filter by any combination of tags, search free text, preview audio, copy file paths, and jump straight to the file in Explorer.
Nothing leaves your machine. No subscriptions. No cloud uploads. No ongoing costs.
What you can search for
– Sound events: `explosion`, `rain`, `footstep`, `piano`, `gunshot`, `thunder` engine
– Category: Sound effect / Music / Ambience / Atmosphere
– Type: One-shot / Short clip / Loop / Full track
– Tempo: Very slow → Very fast
– Energy: Quiet / Medium / Loud
– Mood: Bright / Dark
– Musical key: e.g. A minor, D major
– Duration range (seconds)
– File format: WAV, MP3, FLAC, OGG, AIFF
What’s included
– tagger.py — the AI scanner (runs once, or re-run to add new files)
– app.py — the local web server
– Browser-based search UI with audio preview
– Full documentation site
– requirements.txt and step-by-step setup instructions
System requirements
– Python 3.10+
– Windows, macOS, or Linux
– NVIDIA GPU recommended (CPU works but is slower)
– ~2 GB disk for the AI model (downloaded once on first run)
– Supports WAV, MP3, FLAC, OGG, AIF/AIFF
Performance
On a modern NVIDIA GPU, expect roughly 1–3 seconds per file. A library of 10,000 files takes around 4–8 hours — something you run overnight. Already-processed files are skipped on subsequent runs, so adding new files later is fast.
Who is this for?
– Game developers and sound designers who work with large SFX libraries
– Video editors and filmmakers with music and sound collections
– Music producers who’ve accumulated hundreds of samples and loops
– Anyone who’s spent more than five minutes hunting for a specific sound