e11c3abb9e
- Added DashScope ASR service implementation for real-time streaming. - Updated ASR provider logic to support DashScope alongside existing providers. - Enhanced runtime metadata resolution to include DashScope as a valid ASR provider. - Modified configuration files and documentation to reflect the addition of DashScope. - Introduced tests to validate DashScope integration and ASR service behavior. - Refactored ASR service factory to accommodate new provider options and modes.
60 lines
2.0 KiB
YAML
60 lines
2.0 KiB
YAML
# Agent behavior configuration (safe to edit per profile)
|
|
# This file only controls agent-side behavior (VAD/LLM/TTS/ASR providers).
|
|
# Infra/server/network settings should stay in .env.
|
|
|
|
agent:
|
|
vad:
|
|
type: silero
|
|
model_path: data/vad/silero_vad.onnx
|
|
threshold: 0.5
|
|
min_speech_duration_ms: 100
|
|
eou_threshold_ms: 800
|
|
|
|
llm:
|
|
# provider: openai | openai_compatible | siliconflow
|
|
provider: openai_compatible
|
|
model: deepseek-v3
|
|
temperature: 0.7
|
|
# Required: no fallback. You can still reference env explicitly.
|
|
api_key: your_llm_api_key
|
|
# Optional for OpenAI-compatible endpoints:
|
|
api_url: https://api.qnaigc.com/v1
|
|
|
|
tts:
|
|
# provider: openai_compatible | siliconflow | dashscope
|
|
# dashscope defaults (if omitted):
|
|
# api_url: wss://dashscope.aliyuncs.com/api-ws/v1/realtime
|
|
# model: qwen3-tts-flash-realtime
|
|
# dashscope_mode: commit (engine splits) | server_commit (dashscope splits)
|
|
# note: dashscope_mode/mode is ONLY used when provider=dashscope.
|
|
provider: openai_compatible
|
|
api_key: your_tts_api_key
|
|
api_url: https://api.siliconflow.cn/v1/audio/speech
|
|
model: FunAudioLLM/CosyVoice2-0.5B
|
|
voice: anna
|
|
speed: 1.0
|
|
|
|
asr:
|
|
# provider: buffered | openai_compatible | siliconflow | dashscope
|
|
# dashscope defaults (if omitted):
|
|
# api_url: wss://dashscope.aliyuncs.com/api-ws/v1/realtime
|
|
# model: qwen3-asr-flash-realtime
|
|
# note: dashscope uses streaming ASR mode (chunk-by-chunk).
|
|
provider: openai_compatible
|
|
api_key: you_asr_api_key
|
|
api_url: https://api.siliconflow.cn/v1/audio/transcriptions
|
|
model: FunAudioLLM/SenseVoiceSmall
|
|
interim_interval_ms: 500
|
|
min_audio_ms: 300
|
|
start_min_speech_ms: 160
|
|
pre_speech_ms: 240
|
|
final_tail_ms: 120
|
|
|
|
duplex:
|
|
enabled: true
|
|
system_prompt: You are a helpful, friendly voice assistant. Keep your responses concise and conversational.
|
|
|
|
barge_in:
|
|
min_duration_ms: 200
|
|
silence_tolerance_ms: 60
|