Description

Tired of transcriptions that fill silences with phrases nobody said? Captio transcribes your video and audio on your own computer with the most accurate speech recognition model available. It also detects and throws out invented text: ghost credits, “thanks for watching”, repeated loops and phrases in music or silence.

Already have the script or the lyrics? Paste them into “I have the exact text” mode. Captio only works out when each word is said, and the text stays exactly as you wrote it. It’s perfect for scripted videos, music videos, lyric videos, courses and ads.

Pick the style for where the video will be published (YouTube, Netflix-style professional rules, Reels/TikTok/Shorts, word by word or songs) and switch styles instantly without transcribing again. Review the result in the editor with video preview, fix whatever you want and export.

Features

Offline transcription with Whisper large-v3, accelerated by NVIDIA graphics cards.
Filters against invented text: ghost phrases, repetitions and text in silence or music.
“I have the exact text” mode for scripts and lyrics, with automatic syncing.
6 ready-made styles plus a custom one: characters per line, lines, words, duration and reading speed.
Instant style switching without transcribing again.
Editor with preview: fix texts and timings, merge or delete subtitles.
Batch processing: each subtitle file is saved next to its video.
Exports SRT, WebVTT, TXT, word-by-word SRT and JSON with word timings.
90+ languages with automatic detection and translation into English.
List of names and key terms so they are spelled right.
Interface in English and Spanish.
Private: your files never leave your computer. No minute limits and no usage fees.

Requirements

Windows 10 or 11, 64-bit.
8 GB RAM (16 GB recommended).
NVIDIA graphics card recommended (6 GB of memory or more). Without one it runs on the processor, more slowly.
6 GB free disk space: program and speech model.
Internet only the first time, to download the speech model.

Includes