Underwright - Manuscript Maker Local Offline Audio Video to Text Tool

Local Whisper/Qwen3 ASR/SenseVoice · Real time bilingual subtitles · Subtitle workshop · AI summary

Convert meetings, podcasts, and videos into editable transcripts on macOS and Windows; Real time bilingual subtitles, subtitle workshops (ASS/QC v2/Beautify), AI summarization, and batch automation. Whisper/Qwen3-ASR/SenseVoice run locally, with audio and text remaining on the device throughout the entire process.

From transcription to subtitle workshop, one computer can handle everything

1. Completely offline transcription

Based on three local engines, OpenAI Whisper, Qwen3 ASR, and SenseVoice, the model can be downloaded and transcribed offline in one go. Not uploading or going through the cloud; Chinese, Japanese, and Korean can switch engines according to the scene, and privacy is completely controllable.

1. Full scene input

Drag and drop local audio and video, real-time recording, capture system sound, or paste YouTube links; Podcast subscriptions and calendar meetings can also be joined with just one click. Batch queues and monitoring folders enable automatic transcription.

1. Real time recording and bilingual subtitles

Record microphone or system audio and automatically transcribe it; The floating subtitle window can display both the original and translated text while watching, and real-time translation supports streaming output. I can keep up with meetings, interviews, live broadcasts, and foreign language videos.

1. AI Intelligence Summary

Transform long texts into action items, key insights, and multiple rounds of Q&A. Can access Claude, DeepSeek, Gemini, Tongyi Qianwen OpenRouter、MiniMax， Or the local Olama model.

1. Speaker separation

Automatically identify and label up to 4 speakers. Meeting minutes, interview organization, and multi person podcasts are differentiated by color, and clicking on them will redirect to the corresponding segments.

1. Subtitle Workshop and Burning

Subtitle table and editing timeline, ASS import/export and style template, QC v2 physical examination, Beautify plastic surgery, spell checking, explicit playlist; 11 formats for export, subtitles can be directly burned into the video.

1. Batch and automation

Queue processing dozens of files, or setting up automatic transfer of monitoring folders; Provide CLI, MCP tools, and Webhooks to integrate transcription into scripts, AI clients, and existing workflows.

1. Audio Enhancement

Noise reduction, voice enhancement, telephone audio optimization, and RNN deep noise reduction preset are also provided; Save low-quality recordings and improve recognition accuracy.

1. Intelligent Chapters and Mind Maps

AI automatically analyzes content structure, generates chapter segmentation and foldable hierarchical mind maps. Long webinars, courses, and presentations can be summarized in just a few seconds.

1. Flexible settings

Choose your own Whisper/Qwen3 ASR/SenseVoice engine, language, and parameters; You can configure 9 UI languages, image sources, and anonymous analysis switches.