agentsclimarketplace

Hub audio

Skill ZhangHanDong/dora-skills/skills/hub-audio

"Use for audio processing nodes in dora.From its SKILL.md

Install
npx -y skills add ZhangHanDong/dora-skills --skill hub-audio

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 7 stars7 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

5.7 KB, as published. Nobody here has run it

Audio Processing Nodes

Microphone input, voice activity detection, speech-to-text, and text-to-speech

Audio Pipeline Overview

Microphone → VAD → Whisper STT → LLM → Kokoro TTS → Speaker

Available Audio Nodes

NodeInstallDescription
dora-microphonepip install dora-microphoneMicrophone input with VAD
dora-vadpip install dora-vadSilero voice activity detection
dora-distil-whisperpip install dora-distil-whisperDistil-Whisper STT
dora-kokoro-ttspip install dora-kokoro-ttsKokoro text-to-speech
dora-pyaudiopip install dora-pyaudioAudio playback

dora-microphone

Capture audio from microphone with built-in voice activity detection.

YAML Configuration

- id: microphone
  build: pip install dora-microphone
  path: dora-microphone
  inputs:
    tick: dora/timer/millis/100
  outputs:
    - audio  # 16kHz Float32Array

Output Format

# audio: Float32Array at 16kHz sample rate
metadata = {"sample_rate": 16000}

dora-vad

Silero Voice Activity Detection - filters audio to speech-only segments.

YAML Configuration

- id: vad
  build: pip install dora-vad
  path: dora-vad
  inputs:
    audio: microphone/audio  # 8kHz or 16kHz
  outputs:
    - audio  # truncated to speech only

Features

  • Detects beginning and ending of voice activity
  • Filters out silence and background noise
  • Maximum voice duration limit to avoid long waits
  • Uses Silero VAD model

dora-distil-whisper

Speech-to-text using Distil-Whisper for efficient transcription.

YAML Configuration

- id: whisper
  build: pip install dora-distil-whisper
  path: dora-distil-whisper
  inputs:
    input: vad/audio
  outputs:
    - text  # StringArray
  env:
    TARGET_LANGUAGE: english  # or other supported languages

Output Format

# text: StringArray containing transcribed text
text = event["value"][0].as_py()  # Get string

dora-kokoro-tts

Efficient text-to-speech using Kokoro.

YAML Configuration

- id: tts
  build: pip install dora-kokoro-tts
  path: dora-kokoro-tts
  inputs:
    text: llm/text
  outputs:
    - audio  # Float32Array

dora-pyaudio

Audio playback through speakers.

YAML Configuration

- id: speaker
  build: pip install dora-pyaudio
  path: dora-pyaudio
  inputs:
    audio: tts/audio

Prerequisites

macOS:

brew install portaudio

Linux:

sudo apt-get install portaudio19-dev python-all-dev

Complete Speech-to-Text Pipeline

nodes:
  # Microphone input
  - id: microphone
    build: pip install dora-microphone
    path: dora-microphone
    inputs:
      tick: dora/timer/millis/100
    outputs:
      - audio

  # Voice activity detection
  - id: vad
    build: pip install dora-vad
    path: dora-vad
    inputs:
      audio: microphone/audio
    outputs:
      - audio

  # Speech to text
  - id: whisper
    build: pip install dora-distil-whisper
    path: dora-distil-whisper
    inputs:
      input: vad/audio
    outputs:
      - text
    env:
      TARGET_LANGUAGE: english

  # Visualization
  - id: rerun
    build: pip install dora-rerun
    path: dora-rerun
    inputs:
      transcription:
        source: whisper/text
        metadata:
          primitive: "text"

Speech-to-Speech Pipeline (Voice Assistant)

nodes:
  # Audio input
  - id: microphone
    build: pip install dora-microphone
    path: dora-microphone
    inputs:
      tick: dora/timer/millis/100
    outputs:
      - audio

  # VAD filtering
  - id: vad
    build: pip install dora-vad
    path: dora-vad
    inputs:
      audio: microphone/audio
    outputs:
      - audio

  # Speech to text
  - id: whisper
    build: pip install dora-distil-whisper
    path: dora-distil-whisper
    inputs:
      input: vad/audio
    outputs:
      - text

  # LLM processing
  - id: llm
    build: pip install dora-qwen
    path: dora-qwen
    inputs:
      text: whisper/text
    outputs:
      - text

  # Text to speech
  - id: tts
    build: pip install dora-kokoro-tts
    path: dora-kokoro-tts
    inputs:
      text: llm/text
    outputs:
      - audio

  # Audio output
  - id: speaker
    build: pip install dora-pyaudio
    path: dora-pyaudio
    inputs:
      audio: tts/audio

Audio Data Format

Float32 Audio Array

import pyarrow as pa
import numpy as np

# Audio at 16kHz
sample_rate = 16000
audio_samples = np.array([...], dtype=np.float32)

# Send audio
audio_data = pa.array(audio_samples)
node.send_output("audio", audio_data, {"sample_rate": sample_rate})

Receiving Audio

audio = event["value"].to_numpy()
sample_rate = event["metadata"].get("sample_rate", 16000)

Troubleshooting

No audio input

# List audio devices
python -c "import sounddevice; print(sounddevice.query_devices())"

PortAudio error

# macOS
brew install portaudio

# Linux
sudo apt-get install portaudio19-dev

Whisper slow on CPU

  • Use smaller model (tiny, base)
  • Consider dora-funasr for Chinese speech recognition

Related Skills

  • hub-llm - Language models for voice assistants
  • hub-visualization - Rerun text visualization
  • domain-audio - Audio pipeline patterns

Keep looking

Skills are one crate of 610,677. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.