I build tools to make noise, aid in the noise-making process/workflow, and help new folks learn how to make their own noise.

• Making music theory easier to see, hear, and explore
• Building focused tools that remove small, stubborn bits of friction
• Programmatic audio rendering with Python, CoreML, and C++
• Audio for interactive media (games!)
• Developing the foundations and core modules for Tonari Labs

_{Web app is live!}

A chord-progression workbench for piano and guitar, with playable voicings, theory context, and a voice companion for co-writing and learning. Started out as a simple way for me to deal with writers block when it comes up. And well now...

• You can describe a harmony/feeling/vibe in natural language
• You can talk with an AI agent (Hanz Hasher!) to talk through an idea, get improv tips for your scale/progression, and much more
• SCALES n' MODES

_{link goes to web app!}

Evver had an interview in your second language? Tonari Tutor is an AI-powered interview practice platform, built specifically for language learners & bilinguals with realistic, live mock interviews. Upload a resume and optional job description, choose a hiring-manager or technical interviewer, practice across languages, and get structured feedback.

A hands-on, interactive learning template for people interested in getting started with programmatic audio. Work through the notebook at your own pace while learning how to load a synth, send MIDI, render audio, experiment with effects, and hear the results as you go.

A fast terminal search tool for browsing installed Native Instruments expansions by name, genre, style, or product type. It scans what is actually installed, cross-references a curated tags database, and keeps recent picks close at hand—so finding the right expansion does not mean digging through menus.

An MCP server that gives your compatible AI assistants deeper access to ElevenLabs Conversational AI—agents, knowledge bases, conversation transcripts and analysis, voices, settings, and webhooks the official connector does not expose. NOT to be confused with, or connected to the TwelveLabs video platform

An experiment in reverse sound synthesis, machine learning, and regressive patch matching. The idea: from audio reference -> audio embeddings weighed against rendered patches, and be flown directly to settings that get closest to that sound/synth. More to come soon, but you can read about where I'm at with it in my lab notes

Tonari Labs · GitHub · Or check out some of my noise!

_AUDIO4ALL.