Update README.md

README.md

@@ -1,3 +1,103 @@
----
-license: mit
----
+# AutoVol (Tray)
+
+AutoVol is a zero-config Windows tray app that **learns when you turn volume down** and then **automatically ducks volume** when the system audio gets louder than your learned “too loud” levels.
+
+It’s designed to be:
+
+* **Zero-touch**: run it and forget it
+* **Self-healing**: tries multiple ways to measure system loudness (loopback audio → peak meter fallbacks)
+* **Non-destructive**: stores its data in your local profile folder, not in system locations
+
+## What it does
+
+* Monitors “how loud the system output is” (best effort loopback capture; fallback peak meters).
+* When you manually turn volume **down**, AutoVol records the loudness level at that moment.
+* Later, if it detects loudness above your learned threshold, it gently reduces volume.
+* Optional speech gating (if available) can reduce ducking during speech-like audio.
+
+## Requirements
+
+* Windows 10/11 (Windows 11 recommended)
+* No admin required (for normal use)
+
+## Install / Run (Release)
+
+AutoVol is typically shipped as a **folder** (PyInstaller *onedir* build).
+
+1. Download the release zip.
+2. Extract it anywhere (ex: `C:\Apps\AutoVol\`).
+3. Run: `AutoVol.exe`
+4. Look for the **tray icon** (system tray / “hidden icons” area).
+
+> **Important:** If the release contains a folder like `AutoVol\AutoVol.exe`, you must keep the **whole folder** together. Do not move only the `.exe` out by itself.
+
+## Tray Menu
+
+Right-click the tray icon for:
+
+* **Observe only (no changes)** — AutoVol will monitor/learn but won’t set volume
+* **Pause / Resume** — stops AutoVol from making adjustments (can restore to your resting volume)
+* **Reset learning** — clears learned “turn-down” history
+* **Open data folder** — opens the app’s local data folder
+* **Quit** — exits cleanly
+
+## Data / Logs Location
+
+AutoVol stores its runtime data under:
+
+`%LOCALAPPDATA%\AutoVol`
+
+You’ll typically see:
+
+* `autovol.log` — rotating logs
+* `autovol_learn.json` — learned “turn-down” samples
+* `autovol_events.jsonl` — event stream (volume changes, sampler changes, etc.)
+* `autovol_crash.txt` — created only if a crash occurs
+
+## Command-line Options
+
+You can run the exe with options (useful for debugging):
+
+* `--observe`
+  Start in observe-only mode (no volume changes).
+
+* `--force-dryvol`
+  Don’t touch system volume (debug mode; logs/stats still run).
+
+* `--reset`
+  Reset learning before starting.
+
+* `--no-autoinstall`
+  Disable auto-install of missing Python dependencies (mostly relevant when running from source).
+
+Example:
+
+```bash
+AutoVol.exe --observe
+```
+
+## Building from Source (Developer)
+
+Basic steps (example):
+
+1. Create/activate a venv
+2. Install requirements
+3. Run:
+
+```bash
+python autovol_tray.py
+```
+
+If you package with PyInstaller, you’ll usually distribute the **onedir** output: `dist\AutoVol\`
+
+## Troubleshooting
+
+* **No tray icon:** Check the Windows tray “hidden icons” area.
+* **No loudness detection / stuck on fallback:** Some machines/drivers don’t support loopback capture; AutoVol will fall back to peak meters. Check `%LOCALAPPDATA%\AutoVol\autovol.log`.
+* **It’s too aggressive:** Use *Observe only* while it learns more “turn-down” events, or reset learning and re-train.
+
+## License
+
+MIT License. 
+
+