minghe36/haoone-app
新一代 AI 专业字幕软件，剪映字幕、elevenlabs 语音转文本的最佳本地版平替之一，也是加强版。中英转录识别准确率超过 97%，词语音频对齐率 98%，说话人分割与识别准确率 96%，带有最先进的 ASR 开源模型，100+字幕动画（支持导出透明背景的字幕动画）。说话人识别、专业字幕编辑器、命令行工具、Skill，达芬奇字幕插件（含字幕动画插件），PR 字幕插件，本地转录、远程转录、文稿匹配、智能拆行、AI校正、AI 智能热词、翻译、双语字幕、专业字幕编辑器、字幕合成、自定义大模型 API
What it solves
It provides a professional, high-accuracy AI subtitling software that serves as a local alternative to tools like CapCut (剪映) and ElevenLabs. It addresses common issues found in Whisper-based tools, such as poor Chinese transcription accuracy and imprecise timestamps, by using more advanced ASR models and a custom word-level audio alignment algorithm.
How it works
The software integrates several state-of-the-art open-source ASR models (including qwen3-asr) and utilizes a transcription engine built with Rust and C++ for high speed and GPU acceleration. It supports both local and remote transcription, offering features like AI-powered line breaking, smart hot-word replacement, and speaker identification. It also integrates directly with professional video editing software via plugins for DaVinci Resolve and Adobe Premiere Pro.
Who it’s for
Video editors, content creators, and translators who need high-precision subtitles in Chinese, English, and Japanese, and who prefer local processing for privacy or speed.
Highlights
- High Accuracy: Claims 97% accuracy for Chinese/English and 94% for Japanese.
- Word-Level Alignment: Custom algorithm achieving 98% audio-to-word alignment accuracy.
- ** wysokie Speed**: Local transcription of a 3-minute video can be completed in under 20 seconds on a Mac Mini M4.
- Professional Integration: Dedicated plugins for DaVinci Resolve and PR, plus a CLI tool.
- Comprehensive Toolset: Includes AI correction, translation, bilingual subtitles, and 100+ subtitle animations.
- Privacy Focused: Promises not to collect user transcription data or audio/video files.
Related
- Project
- Project
- Project
- Project
- Project