dsh-plugin-voz-copilot

Configuration

The defaults come from the package bundle patch and boot on any machine. To use your own workspace and voice, edit the profile patch (~/.dsh/profiles/web/cordis.patch.yml). A patch under the same id replaces the whole config of the line, so restate every key you want to keep.

- id: voz-copilot
  config:
    mode: 'attach'
    workspace: '/path/to/your/workspace'
    title: 'conversa_copilot'
    statePath: '~/.dsh/voz-copilot.json'
    pollMs: 3000
    whisperUrl: 'http://127.0.0.1:8178'
    whisperFastUrl: 'http://127.0.0.1:8179'
    voz:
      persona: 'Answer in short, plain sentences that sound natural when spoken aloud.'
      maxChars: 220
    tts:
      bin: 'edge-tts'
      voice: 'en-US-AriaNeural'
      rate: '+0%'
      volume: '+0%'
      lexicon: '/path/to/tts_lexicon.py'
      python: 'python3'
      servidorQuente: 'http://127.0.0.1:8180'
      servidorQuenteScript: '/path/to/tts-servidor.py'
      servidorQuentePython: 'python3'
    orador:
      limiar: 0.5
      perfilPath: '~/.dsh/voz-perfil.json'

~ is expanded in workspace, statePath and the other paths below.

Key	Default	Purpose
mode	attach (bundle line); code default is title	attach for the web profile; title for the acp profile
workspace	. (the process working directory)	workspace that owns the voice session; set it to your workspace path
title	conversa_copilot	title pinned on the push-to-talk session
marker	none, required in title mode	first-line marker the bridge sends, e.g. [conversa_copilot]
statePath	~/.dsh/voz-copilot.json	durable link between the bridge and the session id
pollMs	3000	how often attach mode checks the link
whisperUrl	http://127.0.0.1:8178	whisper-server used by dictation and by the slow path
whisperFastUrl	same as whisperUrl	faster whisper-server for conversation segments
voz.persona	a short spoken-answer instruction	prefix prepended to each spoken turn
voz.maxChars	220	cap per spoken phrase (60 to 600)
tts.bin	required with tts	edge-tts executable, or a compatible wrapper
tts.voice	required with tts	voice name passed to edge-tts
tts.rate, tts.volume, tts.pitch	unset	passed through when set
tts.lexicon	unset	pronunciation lexicon script (text in, text out)
tts.python	python3	interpreter for the lexicon
tts.servidorQuente	unset	warm TTS HTTP server base URL
tts.servidorQuenteScript	unset	script the plugin starts for the warm server
tts.servidorQuentePython	python3	interpreter for that script
orador.limiar	0.5	cosine threshold for "this is my voice" (0.1 to 0.95)
orador.perfilPath	~/.dsh/voz-perfil.json	voice profile, mode 0600, vectors only
orador.url	http://127.0.0.1:8181	speaker embedding server
orador.python	venv/bin/python in the package (venv\Scripts\python.exe on Windows)	interpreter for the speaker server
orador.script	speaker-servidor.py in the package	speaker server script
orador.modelo	models/orador_eres2net.onnx in the package	speaker embedding model

Text-to-speech

tts.bin is spawned once per phrase with --voice, --text, and the optional --rate=, --volume=, --pitch=, and must write MP3 to stdout. With tts.servidorQuente set, the plugin first tries GET /fala?texto=...&voz=... on that server and starts tts.servidorQuenteScript when it is down, falling back to the CLI. The warm server is what makes the first byte fast; the CLI works without it.

Speaker verification

The default paths assume a checkout of this repository with a venv inside it:

git clone https://github.com/songjongdevsecops/dsh-plugin-voz-copilot.git
cd dsh-plugin-voz-copilot
python3 -m venv venv
venv/bin/pip install -r requirements.txt

Then install the plugin from the checkout (dsh plugin --profile web add /path/to/dsh-plugin-voz-copilot) so the bundled paths resolve, or point orador.python, orador.script and orador.modelo at that checkout. The Python server is optional: without it the conversation mode still works, just without speaker filtering.