dseditor/QwenASRMiniTool

基於OpenVino-int8權重，精簡的QwenASR小工具，用於即時辨識以及字幕轉換使用

What it solves

QwenASRMiniTool is a local speech-to-text tool that allows users to generate subtitles (SRT) from audio files, videos, and live microphone recordings without their data leaving their computer. It simplifies the process of offline transcription for non-experts, providing a free and private alternative to cloud-based services.

How it works

The tool uses a WebView-based interface and supports multiple inference backends to ensure compatibility across different hardware:

  • Inference Cores: It utilizes CrispASR (via Vulkan GPU) for NVIDIA, AMD, and Intel GPUs, and OpenVINO for pure CPU execution.
  • Models: It leverages models like Qwen3-ASR (including a specialized Japanese anime version) and Whisper (Breeze-ASR-26) for high-accuracy transcription across 30 languages.
  • Processing: It uses FFmpeg for extracting audio from video files and Silero-VAD for voice activity detection to handle pauses during live recording.
  • Additional Features: It includes a ForcedAligner for word-level timestamps (Karaoke mode) and an ONNX-based engine for speaker diarization.

Who it’s for

  • Individuals needing private, offline transcription of meetings, podcasts, or lectures.
  • Content creators who want to generate subtitles for videos or lyrics with word-level precision.
  • Users with varying hardware (from high-end GPUs to basic CPUs) who need a simple, "out-of-the-box" ASR tool.

Highlights

  • Local & Private: All processing happens on the device; no data is uploaded to the cloud.
  • Hardware Agnostic: Supports GPU acceleration via Vulkan (NVIDIA/AMD/Intel) and CPU-only mode via OpenVINO.
  • Karaoke Mode: Provides word-level highlighting for precise subtitle alignment and review.
  • Speaker Diarization: Automatically identifies and labels different speakers in the transcription.
  • Batch Processing: Allows importing and processing multiple audio/video files sequentially.
  • API & Remote Access: Includes an OpenAI-compatible transcription API and a Cloudflare tunnel for secure remote uploads from mobile devices.

Related

  • Project
    Blue-B/WhisperSubTranslate

    A local application that converts video audio into multilingual subtitles using whisper.cpp for transcription and Hy-MT2 or cloud APIs for translation.

  • Project
    Quantatirsk/qwen3-asr

    Qwen3‑ASR is a self‑hosted speech‑to‑text server that wraps Qwen3‑ASR models (0.6 B & 1.7 B). It auto‑selects a GPU‑vLLM backend or a CPU‑Rust backend, offers OpenAI‑ and Alibaba‑compatible HTTP/WebSocket APIs, supports speaker diarization, VAD‑based segmentation, batch inference, and can be deployed via Docker, custom GPU images, or offline tarballs. MIT‑licensed.

  • Project
    chidiwilliams/buzz

    An offline transcription and translation tool powered by OpenAI's Whisper that supports local files, YouTube links, and live audio.

  • Project
    collabora/WhisperLive

    WhisperLive is an open‑source, near‑real‑time transcription server built around OpenAI’s Whisper model. It supports multiple inference back‑ends (faster‑whisper, NVIDIA TensorRT, Intel OpenVINO, AMD ROCm), offers word‑level timestamps, hot‑word boosting, speaker diarization and optional translation, and provides both simple command‑line clients and a Python streaming API. Docker images are provided for GPU and CPU deployments, and browser/iOS extensions let you transcribe audio directly from a web page or mobile device.

  • Project
    royshil/obs-localvocal

    An OBS plugin that provides real-time, local speech-to-text transcription and translation using OpenAI's Whisper, ensuring privacy and no cloud costs.