f0458b9b40
- New DeepgramSTT class using Deepgram nova-3 via REST API - Factory function create_stt_engine() for provider switching - faster-whisper import now optional (graceful fallback) - Config defaults to cloud providers (deepgram STT + venice TTS) - .env.example updated with DEEPGRAM_API_KEY and VENICE_API_KEY - requirements.txt adds deepgram-sdk, marks faster-whisper as optional - Zero GPU required for default configuration
261 lines
7 KiB
YAML
261 lines
7 KiB
YAML
# Jarvis Voice Bot Configuration
|
|
# Environment variables in .env override these values
|
|
|
|
# ============================================================================
|
|
# Discord Settings
|
|
# ============================================================================
|
|
discord:
|
|
# Bot token from Discord Developer Portal
|
|
# REQUIRED: Set via DISCORD_TOKEN environment variable
|
|
token: null
|
|
|
|
# Command prefix for text commands (if needed)
|
|
command_prefix: "/"
|
|
|
|
# Bot status message
|
|
status_message: "Listening in voice channels"
|
|
|
|
# Auto-join voice channel on bot start (if user is in voice)
|
|
auto_join: false
|
|
|
|
# ============================================================================
|
|
# Agent Configuration
|
|
# ============================================================================
|
|
agents:
|
|
# Default agent (jarvis or sage)
|
|
default: "jarvis"
|
|
|
|
# Per-agent settings
|
|
jarvis:
|
|
# TTS voice reference file (relative to server/voices/)
|
|
voice_file: "jarvis.mp3"
|
|
|
|
# Agent personality for LLM context
|
|
personality: |
|
|
You are Jarvis, an intelligent, witty, and helpful AI assistant.
|
|
You speak naturally and conversationally, with subtle British sophistication.
|
|
You provide accurate information and thoughtful insights without being
|
|
verbose. You have a dry sense of humor but know when to be serious.
|
|
|
|
# TTS emotion exaggeration (0.0 = none, 1.0 = full)
|
|
emotion_exaggeration: 0.3
|
|
|
|
sage:
|
|
voice_file: "sage.wav"
|
|
personality: |
|
|
You are Sage, a wise, calm, and philosophical AI assistant.
|
|
You speak thoughtfully and deliberately, offering deep insights and
|
|
perspectives. You are patient, empathetic, and help people think through
|
|
complex problems. Your tone is warm and encouraging.
|
|
emotion_exaggeration: 0.2
|
|
|
|
# ============================================================================
|
|
# OpenClaw Gateway
|
|
# ============================================================================
|
|
openclaw:
|
|
# WebSocket URL for OpenClaw Gateway
|
|
# REQUIRED: Set via OPENCLAW_BASE_URL environment variable
|
|
# Format: ws://IP:PORT (default port: 18789)
|
|
base_url: null
|
|
|
|
# Authentication token
|
|
# REQUIRED: Set via OPENCLAW_AUTH_TOKEN environment variable
|
|
token: null
|
|
|
|
# Request timeout (seconds)
|
|
timeout: 8.0
|
|
|
|
# Retry timeout (seconds)
|
|
retry_timeout: 15.0
|
|
|
|
# Retry attempts on failure
|
|
max_retries: 1
|
|
|
|
# Model/agent selection
|
|
model: "claude-sonnet-4"
|
|
|
|
# Agent ID for session keys
|
|
agent_id: "jarvis"
|
|
|
|
# Session scope: per-peer or shared
|
|
session_scope: "per-peer"
|
|
|
|
# ============================================================================
|
|
# Pipeline Configuration
|
|
# ============================================================================
|
|
pipeline:
|
|
# Voice Activity Detection (Silero VAD)
|
|
vad:
|
|
# Silence duration to consider speech ended (seconds)
|
|
silence_threshold: 0.3
|
|
|
|
# Minimum speech duration to process (seconds)
|
|
min_speech_duration: 0.5
|
|
|
|
# VAD confidence threshold (0.0-1.0)
|
|
speech_threshold: 0.5
|
|
|
|
# Smart Turn v3 Configuration
|
|
turn_detection:
|
|
# Turn completion confidence threshold (0.0-1.0)
|
|
# Higher = more certain turn is complete before proceeding
|
|
threshold: 0.7
|
|
|
|
# Maximum wait time after silence before forcing completion (seconds)
|
|
max_wait: 3.0
|
|
|
|
# Model path (relative to models/ directory)
|
|
# Using v3.2 GPU model for best performance with RTX 5090
|
|
model_path: "smart-turn-v3.2-gpu.onnx"
|
|
|
|
# Speech-to-Text
|
|
stt:
|
|
# Provider: "deepgram" (cloud, no GPU) or "local" (faster-whisper, requires GPU)
|
|
provider: "deepgram"
|
|
|
|
# Deepgram settings (used when provider is "deepgram")
|
|
model: "nova-3"
|
|
language: "en"
|
|
|
|
# Local faster-whisper settings (used when provider is "local")
|
|
model_size: "small"
|
|
device: "cuda"
|
|
compute_type: "float16"
|
|
beam_size: 1
|
|
|
|
# Language hint (null = auto-detect)
|
|
language: "en"
|
|
|
|
# VAD filter (use built-in VAD in whisper)
|
|
vad_filter: false
|
|
|
|
# Relevance Filter
|
|
relevance:
|
|
# Default sensitivity: low, medium, high
|
|
default_sensitivity: "medium"
|
|
|
|
# Sensitivity thresholds (LLM confidence 0.0-1.0)
|
|
thresholds:
|
|
low: 1.0 # Only fast path (name mentions)
|
|
medium: 0.75 # Fast path + LLM with 75% confidence
|
|
high: 0.5 # Fast path + LLM with 50% confidence
|
|
|
|
# LLM for classification (if not using OpenClaw)
|
|
# Can be: openai, anthropic, local, openclaw
|
|
classifier: "openclaw"
|
|
|
|
# Classification timeout (seconds)
|
|
timeout: 2.0
|
|
|
|
# Cache classifications (avoid re-classifying similar utterances)
|
|
enable_cache: true
|
|
cache_ttl: 300 # seconds
|
|
|
|
# Transcript Management
|
|
transcript:
|
|
# Rolling window duration (seconds)
|
|
window_duration: 90
|
|
|
|
# Maximum number of turns to keep
|
|
max_turns: 20
|
|
|
|
# Timezone for timestamp display
|
|
timezone: "America/Los_Angeles"
|
|
|
|
# Text-to-Speech
|
|
tts:
|
|
# Provider: "venice" (cloud, no GPU) or "local" (chatterbox, requires GPU)
|
|
provider: "venice"
|
|
|
|
# Venice settings (used when provider is "venice")
|
|
venice:
|
|
voice: "am_liam"
|
|
base_url: "https://api.venice.ai/api/v1"
|
|
# API key from env: VENICE_API_KEY
|
|
|
|
# Local settings (used when provider is "local")
|
|
engine: "chatterbox"
|
|
device: "cuda"
|
|
|
|
# Streaming: generate and play audio in chunks
|
|
streaming: true
|
|
|
|
# Chunk duration for streaming (seconds)
|
|
chunk_duration: 0.5
|
|
|
|
# Voice cloning settings (for Coqui XTTS)
|
|
coqui:
|
|
model_name: "tts_models/multilingual/multi-dataset/xtts_v2"
|
|
language: "en"
|
|
temperature: 0.75
|
|
length_penalty: 1.0
|
|
repetition_penalty: 5.0
|
|
top_k: 50
|
|
top_p: 0.85
|
|
|
|
# Audio Buffering
|
|
audio:
|
|
# Buffer duration per user (seconds)
|
|
buffer_duration: 10.0
|
|
|
|
# Sample rate for processing (Hz)
|
|
processing_sample_rate: 16000
|
|
|
|
# Discord audio sample rate (Hz)
|
|
discord_sample_rate: 48000
|
|
|
|
# ============================================================================
|
|
# FastAPI Server
|
|
# ============================================================================
|
|
server:
|
|
# Server host
|
|
host: "0.0.0.0"
|
|
|
|
# Server port
|
|
port: 8880
|
|
|
|
# Enable TTS endpoint
|
|
enable_tts: true
|
|
|
|
# Enable STT endpoint
|
|
enable_stt: true
|
|
|
|
# API key for authentication (optional)
|
|
# Set via SERVER_API_KEY environment variable
|
|
api_key: null
|
|
|
|
# CORS settings
|
|
cors:
|
|
enabled: true
|
|
allowed_origins: ["*"]
|
|
allowed_methods: ["*"]
|
|
allowed_headers: ["*"]
|
|
|
|
# ============================================================================
|
|
# Logging
|
|
# ============================================================================
|
|
logging:
|
|
# Log level: DEBUG, INFO, WARNING, ERROR, CRITICAL
|
|
level: "INFO"
|
|
|
|
# Log format
|
|
format: "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
|
|
|
|
# Enable latency tracking
|
|
track_latency: true
|
|
|
|
# Per-module log levels (override global level)
|
|
modules:
|
|
discord_bot: "INFO"
|
|
pipeline: "INFO"
|
|
server: "INFO"
|
|
openclaw_client: "DEBUG"
|
|
|
|
# Log file (optional, null = console only)
|
|
file: null
|
|
|
|
# Rotate logs
|
|
rotation:
|
|
enabled: false
|
|
max_bytes: 10485760 # 10MB
|
|
backup_count: 5
|