BhashaBlend (AKA SynthoTranslate) is a video dubbing tool addressing language and hearing impairments in Indian education. It utilizes NLP, speech processing, and computer vision to provide multilingual content, sign language transcription, and voice cloning and translation for inclusive learning.

Server for running Silero TTS models (or other compatible models) with OpenTTS-like API. This server allows users to generate speech using different models and provides an easy-to-use REST API.

JVM library for noise reduction written in Kotlin based on the ML model Silero

Silero VAD engine implementation extending the SpeechMux VAD plugin base

JVM library for voice activity detection written in Kotlin based on C library fvad and Silero

Доработка ИИ агента для кейса на Весенней экономической школе Сбера и НИУ ВШЭ. Решение заняло 3-е место

Распознание и озвучивание голосовым движком текста с экрана.

Self-hosted Text-to-Speech (TTS) server for OpenClaw

🎤 Speech recognition experiments: Whisper, Wav2Vec, Silero ensemble pipelines

Silero VAD Engram for bobrapet — ONNX-based voice activity detection for streaming audio pipelines.

This is a python project. We compare the metrics of 2 already trained AI models - WebRTC & Silero.

Rust implementation of the Silero Voice Activity Detection (VAD) model

It's typescript based VAD that uses silero ai VAD under the hood. It's highly robust for Voice Activity Detection. It only works in the browser.

Convert PDFs/EPUBs to audiobooks with synchronized text highlighting using AI TTS models

Modular Swift package for on-device voice activity detection on Apple platforms using CoreML and Silero VAD.

Browser audio recorder with Whisper transcription

🔊 Clone voices easily with this lightweight Python app that synthesizes audio using a simple voice-cloning workflow.

Automatically cuts out parts without speech from given video, making it shorter and more enjoyable to watch (look examples). Usage on google.collab in several clicks.

A web-app/library for transcribing speech