NoNoise-Mac AGENTS.md

A set of instructions for working on NoNoise Mac, a macOS menu-bar app that removes microphone background noise and room echo on the device. It explains the project, its users, architecture, and build, run, and test process.

In plain words
What is it for?
Use it when developing, testing, or reviewing the NoNoise Mac codebase, especially its real-time audio processing and macOS integration.
Why use it?
It gives an agent the project context and working rules needed before changing the audio code.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/ivalsaraj/nonoise-mac/agents-md
Clone the repo
git clone --depth 1 https://github.com/ivalsaraj/NoNoise-Mac

Made for: Codex, OpenCode.

Per session 11,917 This file is loaded in full into every session.
When invoked 11,917 The same file — it is already loaded in full.
Security scan B 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.11917 $0.11917
Opus 5 $0.05958 $0.05958
Sonnet 5 $0.02383 $0.02383
Haiku 4.5 $0.01192 $0.01192

Measured 2d ago against content hash 87d9b6d65025, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

NoNoise-Mac AGENTS.md scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Asks for rootmediumPrivilege escalation

A mod that escalates privileges can change anything on the machine, not only the project.

The virtual mic has its own scripts: `./build-driver.sh`, `sudo ./install-driver.sh`,
AGENTS.md · 340 lines

How it starts

The opening of the file, as written. The whole thing — 340 lines — stays where its author put it; the contents beside it link to each section on GitHub.

NoNoise Mac — Agent Operating Guide

This file is the entry point for any AI agent (or human) working in this repo. Read it first, then the linked knowledge base before touching audio code.

What this is

NoNoise Mac is a macOS (Apple Silicon) menu-bar app that removes background noise and room reverb from your microphone in real time, 100% on-device, using DeepFilterNet3 on CoreML + Metal / Accelerate. It routes the cleaned voice through a virtual audio cable so any app (Zoom, Meet, Discord, OBS, …) receives studio-clean audio.

ICP / Who We Build For

  • macOS Apple-Silicon users on calls, podcasts, livestreams, and screen recordings who need clean voice without cloud processing or a subscription.
  • Privacy-conscious users — no audio ever leaves the machine.
  • Creators who want light studio polish (EQ + leveling) without opening a DAW.
  • Not for: Windows/Linux, Intel Macs, or multi-channel music mastering.

Architecture map

  • Sources/Core — engine, no UI:
    • AudioModel — CoreAudio/AVFoundation capture + playback, ring buffer, render callback, preset/knob state + persistence.
    • AudioProcessing/DeepFilterNetDSP — STFT → DeepFilterNet feature pipeline → CoreML call → ISTFT → wet/dry blend.
    • AudioProcessing/DeepFilterNet3_Streaming — generated CoreML model wrapper.
    • AudioProcessing/VoiceChain + Biquad + Dynamics — post-DSP "voice polish" (high-pass → shelves → compressor → limiter) plus the optional Broadcast Voice clarity stages (presence peaking bell → subtractive DeEsser) and optional Mouth Noise finisher stages (subtractive DePlosive → broadband-gate DeClick). Driven by ClarityLevel + MouthNoiseLevel, each is gated independently of the noise preset.
    • AudioProcessing/IncomingCleanupEngine — a SECOND, independent capture→clean→play pipeline ("clean the other side"), macOS 14.4+. Captures all system audio except NoNoise via a Core Audio process tap (no BlackHole/loopback), runs its OWN DeepFilterNetDSP instance (DFN only — no VoiceChain), and re-renders the cleaned audio to the current default output (auto-following device changes; tapped originals muted). NOT an AudioModel (no mic coupling, no auto-route to the NoNoise Mic sink). Held by AudioModel as an OPTIONAL (stored AnyObject?, gated type) created ONLY while enabled. The tap IOProc → AVAudioSourceNode are bridged by the lock-free TapAudioRing / CTapRing SPSC ring. See "Incoming / guest cleanup" below.
    • AudioProcessing/TapAudioRing + Sources/CTapRing (tap_ring.{c,h}) — a lock-free C11-atomics single-producer/single-consumer float FIFO bridging the two realtime threads of the tap-based incoming path. Mirrors the driver's tested nn_ring acquire/release discipline.
    • AudioProcessing/IncomingTapLogic — pure, headless-tested decisions for the incoming path (own-process-object validity, re-pin-vs-rebuild) + the IncomingCleanupStatus enum the UI binds to.
    • AudioProcessing/RingBuffer, SpecHistoryRingBuffer, AudioUtils — buffers + helpers.
    • VoicePreset — preset → DSP params + voice-chain settings (single source of truth).
    • CLIArguments + AudioDenoiseOptions — shared CLI parser (live device, --action, --denoise).
    • AudioFileDenoiser — offline file decode → mono 48 kHz → DeepFilterNetDSP → optional VoiceChain → write WAV/CAF/etc. Waits on DeepFilterNetDSP.waitUntilReady() before processing; writes to a temp sibling then moves on success.
  • Sources/App — SwiftUI menu-bar app: NoNoiseMacApp, ContentView (popover), SettingsView, and LaunchAtLoginManager (macOS Service Management adapter).
  • Sources/CLINoNoiseMacCLI: live device pipeline, one-shot --action URL dispatch, and --denoise offline file mode (audio containers only — no MP4/video remux in v1).
  • ResourcesDeepFilterNet3_Streaming.mlmodelc, AppIcon.icns, NoNoiseMacLogo.png, Info.plist, NoNoiseMac.entitlements.
  • Tests/NoNoiseMacTests — pure DSP / preset / voice-chain unit tests (run headless).

Read the full file on GitHub · 340 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 340 lines · 11,917 tokens per session scan B 87d9b6d65025

Subscribe to this mod's changes

NoNoise-Mac AGENTS.md is an instructions file published in the GitHub repository ivalsaraj/NoNoise-Mac (12 stars, last pushed 1mo ago), licensed MIT. It adds 11,917 tokens to every session, about $0.0596 per session on Opus 5. A static security scan graded it B with 1 finding (asks for root). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other instructions, from other repositories

jive-vocals AGENTS.md

AGENTS.md instructions for linuxmatters/jive-vocals, covering agents.md, project overview, setup commands, build and test commands and architecture.

linuxmatters/jive-vocals · 10,147 tokens

jive-vocals CLAUDE.md

Claude Code instructions for linuxmatters/jive-vocals, a project described as: Raw microphone recordings into broadcast-ready audio in one command. No configuration, and no surprises 🗣️.

linuxmatters/jive-vocals · 5 tokens

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

spec-kit AGENTS.md

AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.

github/spec-kit · 7,104 tokens

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,182 tokens

langchain AGENTS.md

AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.

langchain-ai/langchain · 4,345 tokens