@local/dsh-answer-sound

Readme

# dsh-answer-sound

Agent answer sound effects for the DeepSeek Harness web GUI — a start tone, a done tone, and an error tone that follow the agent's answering lifecycle.

English | [中文](README.zh.md)

## Features

- **Answer lifecycle sounds** — a hidden watcher tracks the conversation's running state: the start tone plays when the agent begins answering, the done tone when it finishes, and the error tone when a turn errors
- **Per-kind settings** — Settings → General gains four rows: a master switch, plus one row per tone kind (start / done / error), each with a volume slider, a custom audio file picker, and a preview button
- **No external resources** — default tones are synthesized with the Web Audio API; optional custom audio files are read locally (never uploaded)
- **Persisted locally** — settings are saved in `localStorage` (key `dsh.answer-sound-effects.config`)

## Install

```sh
dsh plugin --profile web add github:<owner>/dsh-answer-sound
```

Restart `dsh web`, then open **Settings → General** to configure.

## Security

This plugin is browser-only: the node half is a no-op, nothing runs on the host, no network requests are made, and audio files never leave your machine.

## License

MIT