OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Audio & Voice AI

Efficientword Net

Efficientword Net is an open-source Python library for adding few-shot spoken-hotword detection to applications. It uses a Siamese-network-inspired approach and TensorFlow Lite for inference, and supports custom phrases through reference files generated from recorded pronunciations.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Efficientword Net — OpenViST project profile in Audio & Voice AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when an application needs a spoken activation phrase and you can provide recorded examples, install its audio dependencies, and integrate a Python library.

Who it is for

  • Developers adding custom voice activation to applications
  • Teams building hands-free interfaces for assistants, IoT devices, or desktop software

Documented uses

  • Generate a reference file for a custom spoken hotword from WAV recordings
  • Detect one or more hotwords in an audio stream
  • Adjust confidence thresholds and trigger spacing for a detector
Before you choose

Requirements & limits

The repository specifies an Apache-2.0 license.

  • The library requires Python 3.10–3.14 and manually installed PyAudio/PortAudio dependencies.
  • Custom reference generation accepts WAV files only.
  • The repository says detection performs best with three to four hotword samples recorded directly from the user.
  • The relaxation-time setting can reduce repeated triggers from overlapping sliding windows.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Create a custom hotword reference file

Prepare a JSON reference file from recorded pronunciations for use with the detector.

Before you start

  • Python 3.10–3.14
  • PyAudio and its PortAudio dependency installed manually before package installation
  • EfficientWord-Net installed with pip install EfficientWord-Net
  • Four to ten distinct pronunciations of the chosen hotword, saved as WAV files in a folder containing no other files

Steps

  1. Install the library with the command below.

    Documented command
    pip install EfficientWord-Net

  2. Place the WAV recordings in a separate folder.

  3. Run the command below.

    Documented command
    python -m eff_word_net.generate_reference

  4. When prompted, provide the recordings folder and the output folder for the generated _ref.json file.

Expected outcome

A hotword reference JSON file that can be passed to a HotwordDetector instance.

Guide limitations

  • Only WAV files are supported by the reference-generation procedure.
  • This procedure prepares a reference file; microphone detection is a separate stage.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
8 October 2026
Original website title
Ant-Brain/EfficientWord-Net · Official homepage title · retained in its source language
How we review sourcesSuggest a correction