Skip to content

Latest commit

 

History

History
81 lines (51 loc) · 3.37 KB

README.md

File metadata and controls

81 lines (51 loc) · 3.37 KB

HiveMind Audio Binary Protocol Plugin

Extends hivemind-core and integrates with ovos-simple-listener, enabling audio-based communication with advanced features for secure, distributed voice assistant functionality.

🌟 Key Features

  • Audio Stream Handling:
    Accepts encrypted binary audio streams, performing WakeWord detection, Voice Activity Detection (VAD), Speech-to-Text (STT), and Text-to-Speech (TTS) directly on the hivemind-listener instance.
    (Lightweight clients like hivemind-mic-satellite only run a microphone and VAD plugin.)

  • STT Service:
    Provides STT via the hivemind-websocket-client, accepting Base64-encoded audio inputs.

  • TTS Service:
    Provides TTS via the hivemind-websocket-client, returning Base64-encoded audio outputs.

  • Secure Plugin Access:
    Running TTS/STT via HiveMind Listener requires an access key, offering fine-grained access control compared to non-authenticated server plugins.

💡 Tip: hivemind-audio-binary-protocol is a plugin for hivemind-core and is compatible with all existing HiveMind clients.

🚀 Getting Started

Installation

pip install hivemind-audio-binary-protocol

Configuration

In your hivemind server.json set "binary_protocol" module to "hivemind-audio-binary-protocol-plugin".

💡 server.json documentation here

configure the plugins to be used like you would in OVOS

  "binary_protocol": {"module": "hivemind-audio-binary-protocol-plugin",
                      "hivemind-audio-binary-protocol-plugin": {
                          "stt": {"module": "XXX-plugin", "XXX-plugin":{}},
                          "tts": {"module": "XXX-plugin", "XXX-plugin":{}},
                          "vad": {"module": "XXX-plugin", "XXX-plugin":{}},
                          "wake_word": "hey_mycroft",
                          "hotwords": {
                              "hey_mycroft": {
                                  "module": "ovos-ww-plugin-precise-lite",
                                  "model": "https://github.com/OpenVoiceOS/precise-lite-models/raw/master/wakewords/en/hey_mycroft.tflite"
                              }
                          }
                      }},

If installed and configured correctly the audio binary protocol will be used when you run hivemind-core listen

🌐 Example Use Cases

  1. Microphone Satellite: Use hivemind-mic-satellite to stream raw audio to the hivemind-listener.

    Microphones handle audio capture and VAD, while the Listener manages WakeWord, STT, and TTS processing.

  2. Authenticated STT/TTS Services: Connect clients securely using access keys for transcribing or synthesizing audio via the HiveMind Listener, ensuring robust access control.

🤝 Contributing

We welcome contributions!

⚖️ License

HiveMind Listener is open-source software, licensed under the Apache 2.0 License.